d/freebsd-nq
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
91e953ae9b
freebsd-nq/sys/kern/kern_mac.c

3946 lines
85 KiB
C
Raw Normal View History

Stubs for the TrustedBSD MAC system calls to permit TrustedBSD MAC userland code to operate on kernel's from the main tree. Not much in this file yet. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 02:04:05 +00:00
/*-
* Copyright (c) 1999, 2000, 2001, 2002 Robert N. M. Watson
* Copyright (c) 2001 Ilmar S. Habibulin
Update NAI copyright to 2003, missed in earlier commits and merges.
2003-04-18 19:57:37 +00:00
* Copyright (c) 2001, 2002, 2003 Networks Associates Technology, Inc.
Stubs for the TrustedBSD MAC system calls to permit TrustedBSD MAC userland code to operate on kernel's from the main tree. Not much in this file yet. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 02:04:05 +00:00
* All rights reserved.
*
* This software was developed by Robert Watson and Ilmar Habibulin for the
* TrustedBSD Project.
*
License clarification and wording changes: NAI has approved removal of clause three, and NAI Labs now goes by the name Network Associates Laboratories.
2002-11-04 01:42:39 +00:00
* This software was developed for the FreeBSD Project in part by Network
* Associates Laboratories, the Security Research Division of Network
* Associates, Inc. under DARPA/SPAWAR contract N66001-01-C-8035 ("CBOSS"),
* as part of the DARPA CHATS research program.
Stubs for the TrustedBSD MAC system calls to permit TrustedBSD MAC userland code to operate on kernel's from the main tree. Not much in this file yet. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 02:04:05 +00:00
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*/
Use __FBSDID().
2003-06-11 00:56:59 +00:00
Stubs for the TrustedBSD MAC system calls to permit TrustedBSD MAC userland code to operate on kernel's from the main tree. Not much in this file yet. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 02:04:05 +00:00
/*
* Framework for extensible kernel access control. Kernel and userland
* interface to the framework, policy registration and composition.
*/
Use __FBSDID().
2003-06-11 00:56:59 +00:00
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
Stubs for the TrustedBSD MAC system calls to permit TrustedBSD MAC userland code to operate on kernel's from the main tree. Not much in this file yet. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 02:04:05 +00:00
#include "opt_mac.h"
Remember to include "opt_devfs.h" so we get any relevant changes to NDEVFSINO before we include devfs.h. Spotted by: FlexeLint
2002-10-01 15:24:35 +00:00
#include "opt_devfs.h"
Include file cleanup; mac.h and malloc.h at one point had ordering relationship requirements, and no longer do. Reminded by: bde
2002-08-01 17:47:56 +00:00
Stubs for the TrustedBSD MAC system calls to permit TrustedBSD MAC userland code to operate on kernel's from the main tree. Not much in this file yet. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 02:04:05 +00:00
#include <sys/param.h>
Introduce a condition variable to avoid returning EBUSY when the MAC policy list is busy during a load or unload attempt. We assert no locks held during the cv wait, meaning we should be fairly deadlock-safe. Because of the cv model and busy count, it's possible for a cv waiter waiting for exclusive access to the policy list to be starved by active and long-lived access control/labeling events. For now, we accept that as a necessary tradeoff. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-11-13 15:47:09 +00:00
#include <sys/condvar.h>
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
#include <sys/extattr.h>
Bring in two sets of changes: (1) Permit userland applications to request a change of label atomic with an execve() via mac_execve(). This is required for the SEBSD port of SELinux/FLASK. Attempts to invoke this without MAC compiled in result in ENOSYS, as with all other MAC system calls. Complexity, if desired, is present in policy modules, rather than the framework. (2) Permit policies to have access to both the label of the vnode being executed as well as the interpreter if it's a shell script or related UNIX nonsense. Because we can't hold both vnode locks at the same time, cache the interpreter label. SEBSD relies on this because it supports secure transitioning via shell script executables. Other policies might want to take both labels into account during an integrity or confidentiality decision at execve()-time. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-11-05 17:51:56 +00:00
#include <sys/imgact.h>
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
#include <sys/kernel.h>
#include <sys/lock.h>
Include <sys/malloc.h> instead of depending on namespace pollution 2 layers deep in <sys/proc.h> or <sys/vnode.h>. Removed unused includes. Fixed some printf format errors (1 fatal on i386's; 1 fatal on alphas; 1 not fatal on any supported machine).
2002-09-05 07:02:43 +00:00
#include <sys/malloc.h>
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
#include <sys/mutex.h>
#include <sys/mac.h>
Declare a module service "kernel_mac_support" when MAC support is enabled and the kernel provides the MAC registration and entry point service. Declare a dependency on that module service for any MAC module registered using mac_policy.h. For now, hard code the version as 1, but once we've come up with a versioning policy, we'll move to a #define of some sort. In the mean time, this will prevent loading a MAC module when 'options MAC' isn't present, which (due to a bug in the kernel linker) can result if the MAC module is preloaded via loader.conf. This particular evil recommended by: peter Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI LAbs
2002-08-12 02:00:21 +00:00
#include <sys/module.h>
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
#include <sys/proc.h>
Redesign the externalization APIs from the MAC Framework to the MAC policy modules to improve robustness against C string bugs and vulnerabilities. Following these revisions, all string construction of labels for export to userspace (or elsewhere) is performed using the sbuf API, which prevents the consumer from having to perform laborious and intricate pointer and buffer checks. This substantially simplifies the externalization logic, both at the MAC Framework level, and in individual policies; this becomes especially useful when policies export more complex label data, such as with compartments in Biba and MLS. Bundled in here are some other minor fixes associated with externalization: including avoiding malloc while holding the process mutex in mac_lomac, and hence avoid a failure mode when printing labels during a downgrade operation due to the removal of the M_NOWAIT case. This has been running in the MAC development tree for about three weeks without problems. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-06-23 01:26:34 +00:00
#include <sys/sbuf.h>
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
#include <sys/systm.h>
Stubs for the TrustedBSD MAC system calls to permit TrustedBSD MAC userland code to operate on kernel's from the main tree. Not much in this file yet. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 02:04:05 +00:00
#include <sys/sysproto.h>
#include <sys/sysent.h>
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
#include <sys/vnode.h>
#include <sys/mount.h>
#include <sys/file.h>
#include <sys/namei.h>
#include <sys/socket.h>
#include <sys/pipe.h>
#include <sys/socketvar.h>
#include <sys/sysctl.h>
#include <vm/vm.h>
#include <vm/pmap.h>
#include <vm/vm_map.h>
#include <vm/vm_object.h>
#include <sys/mac_policy.h>
#include <fs/devfs/devfs.h>
#include <net/bpfdesc.h>
#include <net/if.h>
#include <net/if_var.h>
#include <netinet/in.h>
#include <netinet/ip_var.h>
#ifdef MAC
Declare a module service "kernel_mac_support" when MAC support is enabled and the kernel provides the MAC registration and entry point service. Declare a dependency on that module service for any MAC module registered using mac_policy.h. For now, hard code the version as 1, but once we've come up with a versioning policy, we'll move to a #define of some sort. In the mean time, this will prevent loading a MAC module when 'options MAC' isn't present, which (due to a bug in the kernel linker) can result if the MAC module is preloaded via loader.conf. This particular evil recommended by: peter Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI LAbs
2002-08-12 02:00:21 +00:00
/*
* Declare that the kernel provides MAC support, version 1. This permits
* modules to refuse to be loaded if the necessary support isn't present,
* even if it's pre-boot.
*/
MODULE_VERSION(kernel_mac_support, 1);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
SYSCTL_DECL(_security);
SYSCTL_NODE(_security, OID_AUTO, mac, CTLFLAG_RW, 0,
"TrustedBSD MAC policy controls");
Push the debugging obect label counters into security.mac.debug.counters rather than directly under security.mac.debug. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:30:53 +00:00
Rename MAC_MAX_POLICIES to MAC_MAX_SLOTS, since the variables and constants in question refer to the number of label slots, not the maximum number of policies that may be loaded. This should reduce confusion regarding an element in the MAC sysctl MIB, as well as make it more clear what the affect of changing the compile-time constants is. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-08 19:49:42 +00:00
#if MAC_MAX_SLOTS > 32
#error "MAC_MAX_SLOTS too large"
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
#endif
If MAC_MAX_POLICIES isn't defined, don't try to define it, just let the compile fail. MAC_MAX_POLICIES should always be defined, or we have bigger problems at hand. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-20 03:41:09 +00:00
Rename MAC_MAX_POLICIES to MAC_MAX_SLOTS, since the variables and constants in question refer to the number of label slots, not the maximum number of policies that may be loaded. This should reduce confusion regarding an element in the MAC sysctl MIB, as well as make it more clear what the affect of changing the compile-time constants is. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-08 19:49:42 +00:00
static unsigned int mac_max_slots = MAC_MAX_SLOTS;
static unsigned int mac_slot_offsets_free = (1 << MAC_MAX_SLOTS) - 1;
SYSCTL_UINT(_security_mac, OID_AUTO, max_slots, CTLFLAG_RD,
&mac_max_slots, 0, "");
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
Comment describing the semantics of mac_late. Trim trailing whitespace. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-25 20:45:27 +00:00
/*
* Has the kernel started generating labeled objects yet? All read/write
* access to this variable is serialized during the boot process. Following
* the end of serialization, we don't update this flag; no locking.
*/
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
static int mac_late = 0;
Slightly change the semantics of vnode labels for MAC: rather than "refreshing" the label on the vnode before use, just get the label right from inception. For single-label file systems, set the label in the generic VFS getnewvnode() code; for multi-label file systems, leave the labeling up to the file system. With UFS1/2, this means reading the extended attribute during vfs_vget() as the inode is pulled off disk, rather than hitting the extended attributes frequently during operations later, improving performance. This also corrects sematics for shared vnode locks, which were not previously present in the system. This chances the cache coherrency properties WRT out-of-band access to label data, but in an acceptable form. With UFS1, there is a small race condition during automatic extended attribute start -- this is not present with UFS2, and occurs because EAs aren't available at vnode inception. We'll introduce a work around for this shortly. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-26 14:38:24 +00:00
/*
* Warn about EA transactions only the first time they happen.
* Weak coherency, no locking.
*/
static int ea_warn_once = 0;
Move MAC label storage for mbufs into m_tags from the m_pkthdr structure, returning some additional room in the first mbuf in a chain, and avoiding feature-specific contents in the mbuf header. To do this: - Modify mbuf_to_label() to extract the tag, returning NULL if not found. - Introduce mac_init_mbuf_tag() which does most of the work mac_init_mbuf() used to do, except on an m_tag rather than an mbuf. - Scale back mac_init_mbuf() to perform m_tag allocation and invoke mac_init_mbuf_tag(). - Replace mac_destroy_mbuf() with mac_destroy_mbuf_tag(), since m_tag's are now GC'd deep in the m_tag/mbuf code rather than at a higher level when mbufs are directly free()'d. - Add mac_copy_mbuf_tag() to support m_copy_pkthdr() and related notions. - Generally change all references to mbuf labels so that they use mbuf_to_label() rather than &mbuf->m_pkthdr.label. This required no changes in the MAC policies (yay!). - Tweak mbuf release routines to not call mac_destroy_mbuf(), tag destruction takes care of it for us now. - Remove MAC magic from m_copy_pkthdr() and m_move_pkthdr() -- the existing m_tag support does all this for us. Note that we can no longer just zero the m_tag list on the target mbuf, rather, we have to delete the chain because m_tag's will already be hung off freshly allocated mbuf's. - Tweak m_tag copying routines so that if we're copying a MAC m_tag, we don't do a binary copy, rather, we initialize the new storage and do a deep copy of the label. - Remove use of MAC_FLAG_INITIALIZED in a few bizarre places having to do with mbuf header copies previously. - When an mbuf is copied in ip_input(), we no longer need to explicitly copy the label because it will get handled by the m_tag code now. - No longer any weird handling of MAC labels in if_loop.c during header copies. - Add MPC_LOADTIME_FLAG_LABELMBUFS flag to Biba, MLS, mac_test. In mac_test, handle the label==NULL case, since it can be dynamically loaded. In order to improve performance with this change, introduce the notion of "lazy MAC label allocation" -- only allocate m_tag storage for MAC labels if we're running with a policy that uses MAC labels on mbufs. Policies declare this intent by setting the MPC_LOADTIME_FLAG_LABELMBUFS flag in their load-time flags field during declaration. Note: this opens up the possibility of post-boot policy modules getting back NULL slot entries even though they have policy invariants of non-NULL slot entries, as the policy might have been loaded after the mbuf was allocated, leaving the mbuf without label storage. Policies that cannot handle this case must be declared as NOTLATE, or must be modified. - mac_labelmbufs holds the current cumulative status as to whether any policies require mbuf labeling or not. This is updated whenever the active policy set changes by the function mac_policy_updateflags(). The function iterates the list and checks whether any have the flag set. Write access to this variable is protected by the policy list; read access is currently not protected for performance reasons. This might change if it causes problems. - Add MAC_POLICY_LIST_ASSERT_EXCLUSIVE() to permit the flags update function to assert appropriate locks. - This makes allocation in mac_init_mbuf() conditional on the flag. Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 20:39:06 +00:00
/*
* Flag to indicate whether or not we should allocate label storage for
* new mbufs. Since most dynamic policies we currently work with don't
* rely on mbuf labeling, try to avoid paying the cost of mtag allocation
* unless specifically notified of interest. One result of this is
* that if a dynamically loaded policy requests mbuf labels, it must
* be able to deal with a NULL label being returned on any mbufs that
* were already in flight when the policy was loaded. Since the policy
* already has to deal with uninitialized labels, this probably won't
* be a problem. Note: currently no locking. Will this be a problem?
*/
Attempt to simplify #ifdef logic for MAC_ALWAYS_LABEL_MBUF. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-08-01 15:45:14 +00:00
#ifndef MAC_ALWAYS_LABEL_MBUF
Move MAC label storage for mbufs into m_tags from the m_pkthdr structure, returning some additional room in the first mbuf in a chain, and avoiding feature-specific contents in the mbuf header. To do this: - Modify mbuf_to_label() to extract the tag, returning NULL if not found. - Introduce mac_init_mbuf_tag() which does most of the work mac_init_mbuf() used to do, except on an m_tag rather than an mbuf. - Scale back mac_init_mbuf() to perform m_tag allocation and invoke mac_init_mbuf_tag(). - Replace mac_destroy_mbuf() with mac_destroy_mbuf_tag(), since m_tag's are now GC'd deep in the m_tag/mbuf code rather than at a higher level when mbufs are directly free()'d. - Add mac_copy_mbuf_tag() to support m_copy_pkthdr() and related notions. - Generally change all references to mbuf labels so that they use mbuf_to_label() rather than &mbuf->m_pkthdr.label. This required no changes in the MAC policies (yay!). - Tweak mbuf release routines to not call mac_destroy_mbuf(), tag destruction takes care of it for us now. - Remove MAC magic from m_copy_pkthdr() and m_move_pkthdr() -- the existing m_tag support does all this for us. Note that we can no longer just zero the m_tag list on the target mbuf, rather, we have to delete the chain because m_tag's will already be hung off freshly allocated mbuf's. - Tweak m_tag copying routines so that if we're copying a MAC m_tag, we don't do a binary copy, rather, we initialize the new storage and do a deep copy of the label. - Remove use of MAC_FLAG_INITIALIZED in a few bizarre places having to do with mbuf header copies previously. - When an mbuf is copied in ip_input(), we no longer need to explicitly copy the label because it will get handled by the m_tag code now. - No longer any weird handling of MAC labels in if_loop.c during header copies. - Add MPC_LOADTIME_FLAG_LABELMBUFS flag to Biba, MLS, mac_test. In mac_test, handle the label==NULL case, since it can be dynamically loaded. In order to improve performance with this change, introduce the notion of "lazy MAC label allocation" -- only allocate m_tag storage for MAC labels if we're running with a policy that uses MAC labels on mbufs. Policies declare this intent by setting the MPC_LOADTIME_FLAG_LABELMBUFS flag in their load-time flags field during declaration. Note: this opens up the possibility of post-boot policy modules getting back NULL slot entries even though they have policy invariants of non-NULL slot entries, as the policy might have been loaded after the mbuf was allocated, leaving the mbuf without label storage. Policies that cannot handle this case must be declared as NOTLATE, or must be modified. - mac_labelmbufs holds the current cumulative status as to whether any policies require mbuf labeling or not. This is updated whenever the active policy set changes by the function mac_policy_updateflags(). The function iterates the list and checks whether any have the flag set. Write access to this variable is protected by the policy list; read access is currently not protected for performance reasons. This might change if it causes problems. - Add MAC_POLICY_LIST_ASSERT_EXCLUSIVE() to permit the flags update function to assert appropriate locks. - This makes allocation in mac_init_mbuf() conditional on the flag. Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 20:39:06 +00:00
static int mac_labelmbufs = 0;
#endif
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
static int mac_enforce_fs = 1;
SYSCTL_INT(_security_mac, OID_AUTO, enforce_fs, CTLFLAG_RW,
&mac_enforce_fs, 0, "Enforce MAC policy on file system objects");
TUNABLE_INT("security.mac.enforce_fs", &mac_enforce_fs);
Merge kld access control checks from the MAC tree: these access control checks permit policy modules to augment the system policy for permitting kld operations. This permits policies to limit access to kld operations based on credential (and other) properties, as well as to perform checks on the kld being loaded (integrity, etc). Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-11-19 22:12:42 +00:00
static int mac_enforce_kld = 1;
SYSCTL_INT(_security_mac, OID_AUTO, enforce_kld, CTLFLAG_RW,
&mac_enforce_kld, 0, "Enforce MAC policy on kld operations");
TUNABLE_INT("security.mac.enforce_kld", &mac_enforce_kld);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
static int mac_enforce_network = 1;
SYSCTL_INT(_security_mac, OID_AUTO, enforce_network, CTLFLAG_RW,
&mac_enforce_network, 0, "Enforce MAC policy on network packets");
TUNABLE_INT("security.mac.enforce_network", &mac_enforce_network);
At the cost of seeming a little gauche, make use of more traditional alphabetization for mac_enforce_pipe sysctl. Obtained from: TrustedBSD Project Sponsored by: DAPRA, NAI Labs
2002-09-18 02:00:19 +00:00
static int mac_enforce_pipe = 1;
SYSCTL_INT(_security_mac, OID_AUTO, enforce_pipe, CTLFLAG_RW,
&mac_enforce_pipe, 0, "Enforce MAC policy on pipe operations");
Add tunables for the existing sysctl twiddles for pipe and vm enforcement so they can be disabled prior to kernel start. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-09-30 20:50:00 +00:00
TUNABLE_INT("security.mac.enforce_pipe", &mac_enforce_pipe);
At the cost of seeming a little gauche, make use of more traditional alphabetization for mac_enforce_pipe sysctl. Obtained from: TrustedBSD Project Sponsored by: DAPRA, NAI Labs
2002-09-18 02:00:19 +00:00
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
static int mac_enforce_process = 1;
SYSCTL_INT(_security_mac, OID_AUTO, enforce_process, CTLFLAG_RW,
&mac_enforce_process, 0, "Enforce MAC policy on inter-process operations");
TUNABLE_INT("security.mac.enforce_process", &mac_enforce_process);
static int mac_enforce_socket = 1;
SYSCTL_INT(_security_mac, OID_AUTO, enforce_socket, CTLFLAG_RW,
&mac_enforce_socket, 0, "Enforce MAC policy on socket operations");
TUNABLE_INT("security.mac.enforce_socket", &mac_enforce_socket);
Centrally manage enforcement of {reboot,swapon,sysctl} using the mac_enforce_system toggle, rather than several separate toggles. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-27 15:50:49 +00:00
static int mac_enforce_system = 1;
SYSCTL_INT(_security_mac, OID_AUTO, enforce_system, CTLFLAG_RW,
&mac_enforce_system, 0, "Enforce MAC policy on system operations");
TUNABLE_INT("security.mac.enforce_system", &mac_enforce_system);
Implement mac_check_system_sysctl(), a MAC Framework entry point to permit MAC policies to augment the security protections on sysctl() operations. This is not really a wonderful entry point, as we only have access to the MIB of the target sysctl entry, rather than the more useful entry name, but this is sufficient for policies like Biba that wish to use their notions of privilege or integrity to prevent inappropriate sysctl modification. Affects MAC kernels only. Since SYSCTL_LOCK isn't in sysctl.h, just kern_sysctl.c, we can't assert the SYSCTL subsystem lockin the MAC Framework. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-27 07:12:34 +00:00
Centrally manage enforcement of {reboot,swapon,sysctl} using the mac_enforce_system toggle, rather than several separate toggles. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-27 15:50:49 +00:00
static int mac_enforce_vm = 1;
Add a toggle to disable VM enforcement. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-09-18 02:02:08 +00:00
SYSCTL_INT(_security_mac, OID_AUTO, enforce_vm, CTLFLAG_RW,
&mac_enforce_vm, 0, "Enforce MAC policy on vm operations");
Add tunables for the existing sysctl twiddles for pipe and vm enforcement so they can be disabled prior to kernel start. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-09-30 20:50:00 +00:00
TUNABLE_INT("security.mac.enforce_vm", &mac_enforce_vm);
Add a toggle to disable VM enforcement. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-09-18 02:02:08 +00:00
Add security.mac.mmap_revocation, a flag indicating whether we should revoke access to memory maps on a process label change. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-09-09 17:12:24 +00:00
static int mac_mmap_revocation = 1;
SYSCTL_INT(_security_mac, OID_AUTO, mmap_revocation, CTLFLAG_RW,
&mac_mmap_revocation, 0, "Revoke mmap access to files on subject "
"relabel");
Sync to trustedbsd_mac tree: default to sigsegv rather than copy-on-write during a label change resulting in an mmap removal. This is "fail stop" behavior, which is preferred, although it offers slightly less transparency. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-15 02:28:32 +00:00
static int mac_mmap_revocation_via_cow = 0;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
SYSCTL_INT(_security_mac, OID_AUTO, mmap_revocation_via_cow, CTLFLAG_RW,
&mac_mmap_revocation_via_cow, 0, "Revoke mmap access to files via "
"copy-on-write semantics, or by removing all write access");
Wrap maintenance of varios nmac{objectname} counters in MAC_DEBUG so we can avoid the cost of a large number of atomic operations if we're not interested in the object count statistics. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-16 14:21:38 +00:00
#ifdef MAC_DEBUG
Push 'security.mac.debug_label_fallback' behind options MAC_DEBUG. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-10-01 03:24:20 +00:00
SYSCTL_NODE(_security_mac, OID_AUTO, debug, CTLFLAG_RW, 0,
"TrustedBSD MAC debug info");
static int mac_debug_label_fallback = 0;
SYSCTL_INT(_security_mac_debug, OID_AUTO, label_fallback, CTLFLAG_RW,
&mac_debug_label_fallback, 0, "Filesystems should fall back to fs label"
"when label is corrupted.");
TUNABLE_INT("security.mac.debug_label_fallback",
&mac_debug_label_fallback);
Push the debugging obect label counters into security.mac.debug.counters rather than directly under security.mac.debug. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:30:53 +00:00
SYSCTL_NODE(_security_mac_debug, OID_AUTO, counters, CTLFLAG_RW, 0,
"TrustedBSD MAC object counters");
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
static unsigned int nmacmbufs, nmaccreds, nmacifnets, nmacbpfdescs,
nmacsockets, nmacmounts, nmactemp, nmacvnodes, nmacdevfsdirents,
Introduce p_label, extensible security label storage for the MAC framework in struct proc. While the process label is actually stored in the struct ucred pointed to by p_ucred, there is a need for transient storage that may be used when asynchronous (deferred) updates need to be performed on the "real" label for locking reasons. Unlike other label storage, this label has no locking semantics, relying on policies to provide their own protection for the label contents, meaning that a policy leaf mutex may be used, avoiding lock order issues. This permits policies that act based on historical process behavior (such as audit policies, the MAC Framework port of LOMAC, etc) can update process properties even when many existing locks are held without violating the lock order. No currently committed policies implement use of this label storage. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-11-20 15:41:25 +00:00
nmacipqs, nmacpipes, nmacprocs;
Push the debugging obect label counters into security.mac.debug.counters rather than directly under security.mac.debug. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:30:53 +00:00
SYSCTL_UINT(_security_mac_debug_counters, OID_AUTO, mbufs, CTLFLAG_RD,
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
&nmacmbufs, 0, "number of mbufs in use");
Push the debugging obect label counters into security.mac.debug.counters rather than directly under security.mac.debug. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:30:53 +00:00
SYSCTL_UINT(_security_mac_debug_counters, OID_AUTO, creds, CTLFLAG_RD,
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
&nmaccreds, 0, "number of ucreds in use");
Push the debugging obect label counters into security.mac.debug.counters rather than directly under security.mac.debug. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:30:53 +00:00
SYSCTL_UINT(_security_mac_debug_counters, OID_AUTO, ifnets, CTLFLAG_RD,
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
&nmacifnets, 0, "number of ifnets in use");
Push the debugging obect label counters into security.mac.debug.counters rather than directly under security.mac.debug. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:30:53 +00:00
SYSCTL_UINT(_security_mac_debug_counters, OID_AUTO, ipqs, CTLFLAG_RD,
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
&nmacipqs, 0, "number of ipqs in use");
Push the debugging obect label counters into security.mac.debug.counters rather than directly under security.mac.debug. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:30:53 +00:00
SYSCTL_UINT(_security_mac_debug_counters, OID_AUTO, bpfdescs, CTLFLAG_RD,
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
&nmacbpfdescs, 0, "number of bpfdescs in use");
Push the debugging obect label counters into security.mac.debug.counters rather than directly under security.mac.debug. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:30:53 +00:00
SYSCTL_UINT(_security_mac_debug_counters, OID_AUTO, sockets, CTLFLAG_RD,
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
&nmacsockets, 0, "number of sockets in use");
Push the debugging obect label counters into security.mac.debug.counters rather than directly under security.mac.debug. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:30:53 +00:00
SYSCTL_UINT(_security_mac_debug_counters, OID_AUTO, pipes, CTLFLAG_RD,
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
&nmacpipes, 0, "number of pipes in use");
Introduce p_label, extensible security label storage for the MAC framework in struct proc. While the process label is actually stored in the struct ucred pointed to by p_ucred, there is a need for transient storage that may be used when asynchronous (deferred) updates need to be performed on the "real" label for locking reasons. Unlike other label storage, this label has no locking semantics, relying on policies to provide their own protection for the label contents, meaning that a policy leaf mutex may be used, avoiding lock order issues. This permits policies that act based on historical process behavior (such as audit policies, the MAC Framework port of LOMAC, etc) can update process properties even when many existing locks are held without violating the lock order. No currently committed policies implement use of this label storage. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-11-20 15:41:25 +00:00
SYSCTL_UINT(_security_mac_debug_counters, OID_AUTO, procs, CTLFLAG_RD,
&nmacprocs, 0, "number of procs in use");
Push the debugging obect label counters into security.mac.debug.counters rather than directly under security.mac.debug. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:30:53 +00:00
SYSCTL_UINT(_security_mac_debug_counters, OID_AUTO, mounts, CTLFLAG_RD,
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
&nmacmounts, 0, "number of mounts in use");
Push the debugging obect label counters into security.mac.debug.counters rather than directly under security.mac.debug. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:30:53 +00:00
SYSCTL_UINT(_security_mac_debug_counters, OID_AUTO, temp, CTLFLAG_RD,
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
&nmactemp, 0, "number of temporary labels in use");
Push the debugging obect label counters into security.mac.debug.counters rather than directly under security.mac.debug. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:30:53 +00:00
SYSCTL_UINT(_security_mac_debug_counters, OID_AUTO, vnodes, CTLFLAG_RD,
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
&nmacvnodes, 0, "number of vnodes in use");
Push the debugging obect label counters into security.mac.debug.counters rather than directly under security.mac.debug. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:30:53 +00:00
SYSCTL_UINT(_security_mac_debug_counters, OID_AUTO, devfsdirents, CTLFLAG_RD,
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
&nmacdevfsdirents, 0, "number of devfs dirents inuse");
Wrap maintenance of varios nmac{objectname} counters in MAC_DEBUG so we can avoid the cost of a large number of atomic operations if we're not interested in the object count statistics. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-16 14:21:38 +00:00
#endif
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
static int error_select(int error1, int error2);
static int mac_policy_register(struct mac_policy_conf *mpc);
static int mac_policy_unregister(struct mac_policy_conf *mpc);
Sync from MAC tree: break out the single mmap entry point into seperate entry points for each occasion: mac_check_vnode_mmap() Check at initial mapping mac_check_vnode_mprotect() Check at mapping protection change mac_check_vnode_mmap_downgrade() Determine if a mapping downgrade should take place following subject relabel. Implement mmap() and mprotect() entry points for labeled vnode policies. These entry points are currently not hooked up to the VM system in the base tree. These changes improve the consistency of the access control interface and offer more flexibility regarding limiting access to vnode mmaping. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-06 02:46:26 +00:00
static void mac_check_vnode_mmap_downgrade(struct ucred *cred,
struct vnode *vp, int *prot);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
static void mac_cred_mmapped_drop_perms_recurse(struct thread *td,
struct ucred *cred, struct vm_map *map);
Modify label allocation semantics for sockets: pass in soalloc's malloc flags so that we can call malloc with M_NOWAIT if necessary, avoiding potential sleeps while holding mutexes in the TCP syncache code. Similar to the existing support for mbuf label allocation: if we can't allocate all the necessary label store in each policy, we back out the label allocation and fail the socket creation. Sync from MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 21:23:47 +00:00
static void mac_destroy_socket_label(struct label *label);
Slightly change the semantics of vnode labels for MAC: rather than "refreshing" the label on the vnode before use, just get the label right from inception. For single-label file systems, set the label in the generic VFS getnewvnode() code; for multi-label file systems, leave the labeling up to the file system. With UFS1/2, this means reading the extended attribute during vfs_vget() as the inode is pulled off disk, rather than hitting the extended attributes frequently during operations later, improving performance. This also corrects sematics for shared vnode locks, which were not previously present in the system. This chances the cache coherrency properties WRT out-of-band access to label data, but in an acceptable form. With UFS1, there is a small race condition during automatic extended attribute start -- this is not present with UFS2, and occurs because EAs aren't available at vnode inception. We'll introduce a work around for this shortly. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-26 14:38:24 +00:00
static int mac_setlabel_vnode_extattr(struct ucred *cred,
struct vnode *vp, struct label *intlabel);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
MALLOC_DEFINE(M_MACPIPELABEL, "macpipelabel", "MAC labels for pipes");
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
MALLOC_DEFINE(M_MACTEMP, "mactemp", "MAC temporary label storage");
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
/*
Clean up locking for the MAC Framework: (1) Accept that we're now going to use mutexes, so don't attempt to avoid treating them as mutexes. This cleans up locking accessor function names some. (2) Rename variables to _mtx, _cv, _count, simplifying the naming. (3) Add a new form of the _busy() primitive that conditionally makes the list busy: if there are entries on the list, bump the busy count. If there are no entries, don't bump the busy count. Return a boolean indicating whether or not the busy count was bumped. (4) Break mac_policy_list into two lists: one with the same name holding dynamic policies, and a new list, mac_static_policy_list, which holds policies loaded before mac_late and without the unload flag set. The static list may be accessed without holding the busy count, since it can't change at run-time. (5) In general, prefer making the list busy conditionally, meaning we pay only one mutex lock per entry point if all modules are on the static list, rather than two (since we don't have to lower the busy count when we're done with the framework). For systems running just Biba or MLS, this will halve the mutex accesses in the network stack, and may offer a substantial performance benefits. (6) Lay the groundwork for a dynamic-free kernel option which eliminates all locking associated with dynamically loaded or unloaded policies, for pre-configured systems requiring maximum performance but less run-time flexibility. These changes have been running for a few weeks on MAC development branch systems. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-07 17:49:24 +00:00
* mac_static_policy_list holds a list of policy modules that are not
* loaded while the system is "live", and cannot be unloaded. These
* policies can be invoked without holding the busy count.
*
* mac_policy_list stores the list of dynamic policies. A busy count is
Introduce a condition variable to avoid returning EBUSY when the MAC policy list is busy during a load or unload attempt. We assert no locks held during the cv wait, meaning we should be fairly deadlock-safe. Because of the cv model and busy count, it's possible for a cv waiter waiting for exclusive access to the policy list to be starved by active and long-lived access control/labeling events. For now, we accept that as a necessary tradeoff. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-11-13 15:47:09 +00:00
* maintained for the list, stored in mac_policy_busy. The busy count
Clean up locking for the MAC Framework: (1) Accept that we're now going to use mutexes, so don't attempt to avoid treating them as mutexes. This cleans up locking accessor function names some. (2) Rename variables to _mtx, _cv, _count, simplifying the naming. (3) Add a new form of the _busy() primitive that conditionally makes the list busy: if there are entries on the list, bump the busy count. If there are no entries, don't bump the busy count. Return a boolean indicating whether or not the busy count was bumped. (4) Break mac_policy_list into two lists: one with the same name holding dynamic policies, and a new list, mac_static_policy_list, which holds policies loaded before mac_late and without the unload flag set. The static list may be accessed without holding the busy count, since it can't change at run-time. (5) In general, prefer making the list busy conditionally, meaning we pay only one mutex lock per entry point if all modules are on the static list, rather than two (since we don't have to lower the busy count when we're done with the framework). For systems running just Biba or MLS, this will halve the mutex accesses in the network stack, and may offer a substantial performance benefits. (6) Lay the groundwork for a dynamic-free kernel option which eliminates all locking associated with dynamically loaded or unloaded policies, for pre-configured systems requiring maximum performance but less run-time flexibility. These changes have been running for a few weeks on MAC development branch systems. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-07 17:49:24 +00:00
* is protected by mac_policy_mtx; the list may be modified only
Introduce a condition variable to avoid returning EBUSY when the MAC policy list is busy during a load or unload attempt. We assert no locks held during the cv wait, meaning we should be fairly deadlock-safe. Because of the cv model and busy count, it's possible for a cv waiter waiting for exclusive access to the policy list to be starved by active and long-lived access control/labeling events. For now, we accept that as a necessary tradeoff. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-11-13 15:47:09 +00:00
* while the busy count is 0, requiring that the lock be held to
* prevent new references to the list from being acquired. For almost
* all operations, incrementing the busy count is sufficient to
* guarantee consistency, as the list cannot be modified while the
* busy count is elevated. For a few special operations involving a
Clean up locking for the MAC Framework: (1) Accept that we're now going to use mutexes, so don't attempt to avoid treating them as mutexes. This cleans up locking accessor function names some. (2) Rename variables to _mtx, _cv, _count, simplifying the naming. (3) Add a new form of the _busy() primitive that conditionally makes the list busy: if there are entries on the list, bump the busy count. If there are no entries, don't bump the busy count. Return a boolean indicating whether or not the busy count was bumped. (4) Break mac_policy_list into two lists: one with the same name holding dynamic policies, and a new list, mac_static_policy_list, which holds policies loaded before mac_late and without the unload flag set. The static list may be accessed without holding the busy count, since it can't change at run-time. (5) In general, prefer making the list busy conditionally, meaning we pay only one mutex lock per entry point if all modules are on the static list, rather than two (since we don't have to lower the busy count when we're done with the framework). For systems running just Biba or MLS, this will halve the mutex accesses in the network stack, and may offer a substantial performance benefits. (6) Lay the groundwork for a dynamic-free kernel option which eliminates all locking associated with dynamically loaded or unloaded policies, for pre-configured systems requiring maximum performance but less run-time flexibility. These changes have been running for a few weeks on MAC development branch systems. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-07 17:49:24 +00:00
* change to the list of active policies, the mtx itself must be held.
* A condition variable, mac_policy_cv, is used to signal potential
* exclusive consumers that they should try to acquire the lock if a
* first attempt at exclusive access fails.
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
*/
Clean up locking for the MAC Framework: (1) Accept that we're now going to use mutexes, so don't attempt to avoid treating them as mutexes. This cleans up locking accessor function names some. (2) Rename variables to _mtx, _cv, _count, simplifying the naming. (3) Add a new form of the _busy() primitive that conditionally makes the list busy: if there are entries on the list, bump the busy count. If there are no entries, don't bump the busy count. Return a boolean indicating whether or not the busy count was bumped. (4) Break mac_policy_list into two lists: one with the same name holding dynamic policies, and a new list, mac_static_policy_list, which holds policies loaded before mac_late and without the unload flag set. The static list may be accessed without holding the busy count, since it can't change at run-time. (5) In general, prefer making the list busy conditionally, meaning we pay only one mutex lock per entry point if all modules are on the static list, rather than two (since we don't have to lower the busy count when we're done with the framework). For systems running just Biba or MLS, this will halve the mutex accesses in the network stack, and may offer a substantial performance benefits. (6) Lay the groundwork for a dynamic-free kernel option which eliminates all locking associated with dynamically loaded or unloaded policies, for pre-configured systems requiring maximum performance but less run-time flexibility. These changes have been running for a few weeks on MAC development branch systems. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-07 17:49:24 +00:00
static struct mtx mac_policy_mtx;
static struct cv mac_policy_cv;
static int mac_policy_count;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
static LIST_HEAD(, mac_policy_conf) mac_policy_list;
Clean up locking for the MAC Framework: (1) Accept that we're now going to use mutexes, so don't attempt to avoid treating them as mutexes. This cleans up locking accessor function names some. (2) Rename variables to _mtx, _cv, _count, simplifying the naming. (3) Add a new form of the _busy() primitive that conditionally makes the list busy: if there are entries on the list, bump the busy count. If there are no entries, don't bump the busy count. Return a boolean indicating whether or not the busy count was bumped. (4) Break mac_policy_list into two lists: one with the same name holding dynamic policies, and a new list, mac_static_policy_list, which holds policies loaded before mac_late and without the unload flag set. The static list may be accessed without holding the busy count, since it can't change at run-time. (5) In general, prefer making the list busy conditionally, meaning we pay only one mutex lock per entry point if all modules are on the static list, rather than two (since we don't have to lower the busy count when we're done with the framework). For systems running just Biba or MLS, this will halve the mutex accesses in the network stack, and may offer a substantial performance benefits. (6) Lay the groundwork for a dynamic-free kernel option which eliminates all locking associated with dynamically loaded or unloaded policies, for pre-configured systems requiring maximum performance but less run-time flexibility. These changes have been running for a few weeks on MAC development branch systems. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-07 17:49:24 +00:00
static LIST_HEAD(, mac_policy_conf) mac_static_policy_list;
Introduce a condition variable to avoid returning EBUSY when the MAC policy list is busy during a load or unload attempt. We assert no locks held during the cv wait, meaning we should be fairly deadlock-safe. Because of the cv model and busy count, it's possible for a cv waiter waiting for exclusive access to the policy list to be starved by active and long-lived access control/labeling events. For now, we accept that as a necessary tradeoff. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-11-13 15:47:09 +00:00
/*
Replace calls to WITNESS_SLEEP() and witness_list() with equivalent calls to WITNESS_WARN().
2003-03-04 21:03:05 +00:00
* We manually invoke WITNESS_WARN() to allow Witness to generate
Introduce a condition variable to avoid returning EBUSY when the MAC policy list is busy during a load or unload attempt. We assert no locks held during the cv wait, meaning we should be fairly deadlock-safe. Because of the cv model and busy count, it's possible for a cv waiter waiting for exclusive access to the policy list to be starved by active and long-lived access control/labeling events. For now, we accept that as a necessary tradeoff. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-11-13 15:47:09 +00:00
* warnings even if we don't end up ever triggering the wait at
* run-time. The consumer of the exclusive interface must not hold
* any locks (other than potentially Giant) since we may sleep for
* long (potentially indefinite) periods of time waiting for the
* framework to become quiescent so that a policy list change may
* be made.
*/
Clean up locking for the MAC Framework: (1) Accept that we're now going to use mutexes, so don't attempt to avoid treating them as mutexes. This cleans up locking accessor function names some. (2) Rename variables to _mtx, _cv, _count, simplifying the naming. (3) Add a new form of the _busy() primitive that conditionally makes the list busy: if there are entries on the list, bump the busy count. If there are no entries, don't bump the busy count. Return a boolean indicating whether or not the busy count was bumped. (4) Break mac_policy_list into two lists: one with the same name holding dynamic policies, and a new list, mac_static_policy_list, which holds policies loaded before mac_late and without the unload flag set. The static list may be accessed without holding the busy count, since it can't change at run-time. (5) In general, prefer making the list busy conditionally, meaning we pay only one mutex lock per entry point if all modules are on the static list, rather than two (since we don't have to lower the busy count when we're done with the framework). For systems running just Biba or MLS, this will halve the mutex accesses in the network stack, and may offer a substantial performance benefits. (6) Lay the groundwork for a dynamic-free kernel option which eliminates all locking associated with dynamically loaded or unloaded policies, for pre-configured systems requiring maximum performance but less run-time flexibility. These changes have been running for a few weeks on MAC development branch systems. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-07 17:49:24 +00:00
static __inline void
mac_policy_grab_exclusive(void)
{
WITNESS_WARN(WARN_GIANTOK | WARN_SLEEPOK, NULL,
"mac_policy_grab_exclusive() at %s:%d", __FILE__, __LINE__);
mtx_lock(&mac_policy_mtx);
while (mac_policy_count != 0)
cv_wait(&mac_policy_cv, &mac_policy_mtx);
}
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
Clean up locking for the MAC Framework: (1) Accept that we're now going to use mutexes, so don't attempt to avoid treating them as mutexes. This cleans up locking accessor function names some. (2) Rename variables to _mtx, _cv, _count, simplifying the naming. (3) Add a new form of the _busy() primitive that conditionally makes the list busy: if there are entries on the list, bump the busy count. If there are no entries, don't bump the busy count. Return a boolean indicating whether or not the busy count was bumped. (4) Break mac_policy_list into two lists: one with the same name holding dynamic policies, and a new list, mac_static_policy_list, which holds policies loaded before mac_late and without the unload flag set. The static list may be accessed without holding the busy count, since it can't change at run-time. (5) In general, prefer making the list busy conditionally, meaning we pay only one mutex lock per entry point if all modules are on the static list, rather than two (since we don't have to lower the busy count when we're done with the framework). For systems running just Biba or MLS, this will halve the mutex accesses in the network stack, and may offer a substantial performance benefits. (6) Lay the groundwork for a dynamic-free kernel option which eliminates all locking associated with dynamically loaded or unloaded policies, for pre-configured systems requiring maximum performance but less run-time flexibility. These changes have been running for a few weeks on MAC development branch systems. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-07 17:49:24 +00:00
static __inline void
mac_policy_assert_exclusive(void)
{
mtx_assert(&mac_policy_mtx, MA_OWNED);
KASSERT(mac_policy_count == 0,
("mac_policy_assert_exclusive(): not exclusive"));
}
Move MAC label storage for mbufs into m_tags from the m_pkthdr structure, returning some additional room in the first mbuf in a chain, and avoiding feature-specific contents in the mbuf header. To do this: - Modify mbuf_to_label() to extract the tag, returning NULL if not found. - Introduce mac_init_mbuf_tag() which does most of the work mac_init_mbuf() used to do, except on an m_tag rather than an mbuf. - Scale back mac_init_mbuf() to perform m_tag allocation and invoke mac_init_mbuf_tag(). - Replace mac_destroy_mbuf() with mac_destroy_mbuf_tag(), since m_tag's are now GC'd deep in the m_tag/mbuf code rather than at a higher level when mbufs are directly free()'d. - Add mac_copy_mbuf_tag() to support m_copy_pkthdr() and related notions. - Generally change all references to mbuf labels so that they use mbuf_to_label() rather than &mbuf->m_pkthdr.label. This required no changes in the MAC policies (yay!). - Tweak mbuf release routines to not call mac_destroy_mbuf(), tag destruction takes care of it for us now. - Remove MAC magic from m_copy_pkthdr() and m_move_pkthdr() -- the existing m_tag support does all this for us. Note that we can no longer just zero the m_tag list on the target mbuf, rather, we have to delete the chain because m_tag's will already be hung off freshly allocated mbuf's. - Tweak m_tag copying routines so that if we're copying a MAC m_tag, we don't do a binary copy, rather, we initialize the new storage and do a deep copy of the label. - Remove use of MAC_FLAG_INITIALIZED in a few bizarre places having to do with mbuf header copies previously. - When an mbuf is copied in ip_input(), we no longer need to explicitly copy the label because it will get handled by the m_tag code now. - No longer any weird handling of MAC labels in if_loop.c during header copies. - Add MPC_LOADTIME_FLAG_LABELMBUFS flag to Biba, MLS, mac_test. In mac_test, handle the label==NULL case, since it can be dynamically loaded. In order to improve performance with this change, introduce the notion of "lazy MAC label allocation" -- only allocate m_tag storage for MAC labels if we're running with a policy that uses MAC labels on mbufs. Policies declare this intent by setting the MPC_LOADTIME_FLAG_LABELMBUFS flag in their load-time flags field during declaration. Note: this opens up the possibility of post-boot policy modules getting back NULL slot entries even though they have policy invariants of non-NULL slot entries, as the policy might have been loaded after the mbuf was allocated, leaving the mbuf without label storage. Policies that cannot handle this case must be declared as NOTLATE, or must be modified. - mac_labelmbufs holds the current cumulative status as to whether any policies require mbuf labeling or not. This is updated whenever the active policy set changes by the function mac_policy_updateflags(). The function iterates the list and checks whether any have the flag set. Write access to this variable is protected by the policy list; read access is currently not protected for performance reasons. This might change if it causes problems. - Add MAC_POLICY_LIST_ASSERT_EXCLUSIVE() to permit the flags update function to assert appropriate locks. - This makes allocation in mac_init_mbuf() conditional on the flag. Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 20:39:06 +00:00
Clean up locking for the MAC Framework: (1) Accept that we're now going to use mutexes, so don't attempt to avoid treating them as mutexes. This cleans up locking accessor function names some. (2) Rename variables to _mtx, _cv, _count, simplifying the naming. (3) Add a new form of the _busy() primitive that conditionally makes the list busy: if there are entries on the list, bump the busy count. If there are no entries, don't bump the busy count. Return a boolean indicating whether or not the busy count was bumped. (4) Break mac_policy_list into two lists: one with the same name holding dynamic policies, and a new list, mac_static_policy_list, which holds policies loaded before mac_late and without the unload flag set. The static list may be accessed without holding the busy count, since it can't change at run-time. (5) In general, prefer making the list busy conditionally, meaning we pay only one mutex lock per entry point if all modules are on the static list, rather than two (since we don't have to lower the busy count when we're done with the framework). For systems running just Biba or MLS, this will halve the mutex accesses in the network stack, and may offer a substantial performance benefits. (6) Lay the groundwork for a dynamic-free kernel option which eliminates all locking associated with dynamically loaded or unloaded policies, for pre-configured systems requiring maximum performance but less run-time flexibility. These changes have been running for a few weeks on MAC development branch systems. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-07 17:49:24 +00:00
static __inline void
mac_policy_release_exclusive(void)
{
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
Clean up locking for the MAC Framework: (1) Accept that we're now going to use mutexes, so don't attempt to avoid treating them as mutexes. This cleans up locking accessor function names some. (2) Rename variables to _mtx, _cv, _count, simplifying the naming. (3) Add a new form of the _busy() primitive that conditionally makes the list busy: if there are entries on the list, bump the busy count. If there are no entries, don't bump the busy count. Return a boolean indicating whether or not the busy count was bumped. (4) Break mac_policy_list into two lists: one with the same name holding dynamic policies, and a new list, mac_static_policy_list, which holds policies loaded before mac_late and without the unload flag set. The static list may be accessed without holding the busy count, since it can't change at run-time. (5) In general, prefer making the list busy conditionally, meaning we pay only one mutex lock per entry point if all modules are on the static list, rather than two (since we don't have to lower the busy count when we're done with the framework). For systems running just Biba or MLS, this will halve the mutex accesses in the network stack, and may offer a substantial performance benefits. (6) Lay the groundwork for a dynamic-free kernel option which eliminates all locking associated with dynamically loaded or unloaded policies, for pre-configured systems requiring maximum performance but less run-time flexibility. These changes have been running for a few weeks on MAC development branch systems. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-07 17:49:24 +00:00
KASSERT(mac_policy_count == 0,
("mac_policy_release_exclusive(): not exclusive"));
mtx_unlock(&mac_policy_mtx);
cv_signal(&mac_policy_cv);
}
static __inline void
mac_policy_list_busy(void)
{
mtx_lock(&mac_policy_mtx);
mac_policy_count++;
mtx_unlock(&mac_policy_mtx);
}
static __inline int
mac_policy_list_conditional_busy(void)
{
int ret;
mtx_lock(&mac_policy_mtx);
if (!LIST_EMPTY(&mac_policy_list)) {
mac_policy_count++;
ret = 1;
} else
ret = 0;
mtx_unlock(&mac_policy_mtx);
return (ret);
}
static __inline void
mac_policy_list_unbusy(void)
{
mtx_lock(&mac_policy_mtx);
mac_policy_count--;
KASSERT(mac_policy_count >= 0, ("MAC_POLICY_LIST_LOCK"));
if (mac_policy_count == 0)
cv_signal(&mac_policy_cv);
mtx_unlock(&mac_policy_mtx);
}
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
/*
* MAC_CHECK performs the designated check by walking the policy
* module list and checking with each as to how it feels about the
* request. Note that it returns its value via 'error' in the scope
* of the caller.
*/
#define MAC_CHECK(check, args...) do { \
struct mac_policy_conf *mpc; \
Clean up locking for the MAC Framework: (1) Accept that we're now going to use mutexes, so don't attempt to avoid treating them as mutexes. This cleans up locking accessor function names some. (2) Rename variables to _mtx, _cv, _count, simplifying the naming. (3) Add a new form of the _busy() primitive that conditionally makes the list busy: if there are entries on the list, bump the busy count. If there are no entries, don't bump the busy count. Return a boolean indicating whether or not the busy count was bumped. (4) Break mac_policy_list into two lists: one with the same name holding dynamic policies, and a new list, mac_static_policy_list, which holds policies loaded before mac_late and without the unload flag set. The static list may be accessed without holding the busy count, since it can't change at run-time. (5) In general, prefer making the list busy conditionally, meaning we pay only one mutex lock per entry point if all modules are on the static list, rather than two (since we don't have to lower the busy count when we're done with the framework). For systems running just Biba or MLS, this will halve the mutex accesses in the network stack, and may offer a substantial performance benefits. (6) Lay the groundwork for a dynamic-free kernel option which eliminates all locking associated with dynamically loaded or unloaded policies, for pre-configured systems requiring maximum performance but less run-time flexibility. These changes have been running for a few weeks on MAC development branch systems. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-07 17:49:24 +00:00
int entrycount; \
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
\
error = 0; \
Clean up locking for the MAC Framework: (1) Accept that we're now going to use mutexes, so don't attempt to avoid treating them as mutexes. This cleans up locking accessor function names some. (2) Rename variables to _mtx, _cv, _count, simplifying the naming. (3) Add a new form of the _busy() primitive that conditionally makes the list busy: if there are entries on the list, bump the busy count. If there are no entries, don't bump the busy count. Return a boolean indicating whether or not the busy count was bumped. (4) Break mac_policy_list into two lists: one with the same name holding dynamic policies, and a new list, mac_static_policy_list, which holds policies loaded before mac_late and without the unload flag set. The static list may be accessed without holding the busy count, since it can't change at run-time. (5) In general, prefer making the list busy conditionally, meaning we pay only one mutex lock per entry point if all modules are on the static list, rather than two (since we don't have to lower the busy count when we're done with the framework). For systems running just Biba or MLS, this will halve the mutex accesses in the network stack, and may offer a substantial performance benefits. (6) Lay the groundwork for a dynamic-free kernel option which eliminates all locking associated with dynamically loaded or unloaded policies, for pre-configured systems requiring maximum performance but less run-time flexibility. These changes have been running for a few weeks on MAC development branch systems. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-07 17:49:24 +00:00
LIST_FOREACH(mpc, &mac_static_policy_list, mpc_list) { \
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
if (mpc->mpc_ops->mpo_ ## check != NULL) \
error = error_select( \
mpc->mpc_ops->mpo_ ## check (args), \
error); \
} \
Clean up locking for the MAC Framework: (1) Accept that we're now going to use mutexes, so don't attempt to avoid treating them as mutexes. This cleans up locking accessor function names some. (2) Rename variables to _mtx, _cv, _count, simplifying the naming. (3) Add a new form of the _busy() primitive that conditionally makes the list busy: if there are entries on the list, bump the busy count. If there are no entries, don't bump the busy count. Return a boolean indicating whether or not the busy count was bumped. (4) Break mac_policy_list into two lists: one with the same name holding dynamic policies, and a new list, mac_static_policy_list, which holds policies loaded before mac_late and without the unload flag set. The static list may be accessed without holding the busy count, since it can't change at run-time. (5) In general, prefer making the list busy conditionally, meaning we pay only one mutex lock per entry point if all modules are on the static list, rather than two (since we don't have to lower the busy count when we're done with the framework). For systems running just Biba or MLS, this will halve the mutex accesses in the network stack, and may offer a substantial performance benefits. (6) Lay the groundwork for a dynamic-free kernel option which eliminates all locking associated with dynamically loaded or unloaded policies, for pre-configured systems requiring maximum performance but less run-time flexibility. These changes have been running for a few weeks on MAC development branch systems. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-07 17:49:24 +00:00
if ((entrycount = mac_policy_list_conditional_busy()) != 0) { \
LIST_FOREACH(mpc, &mac_policy_list, mpc_list) { \
if (mpc->mpc_ops->mpo_ ## check != NULL) \
error = error_select( \
mpc->mpc_ops->mpo_ ## check (args), \
error); \
} \
mac_policy_list_unbusy(); \
} \
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
} while (0)
/*
* MAC_BOOLEAN performs the designated boolean composition by walking
* the module list, invoking each instance of the operation, and
* combining the results using the passed C operator. Note that it
* returns its value via 'result' in the scope of the caller, which
* should be initialized by the caller in a meaningful way to get
* a meaningful result.
*/
#define MAC_BOOLEAN(operation, composition, args...) do { \
struct mac_policy_conf *mpc; \
Clean up locking for the MAC Framework: (1) Accept that we're now going to use mutexes, so don't attempt to avoid treating them as mutexes. This cleans up locking accessor function names some. (2) Rename variables to _mtx, _cv, _count, simplifying the naming. (3) Add a new form of the _busy() primitive that conditionally makes the list busy: if there are entries on the list, bump the busy count. If there are no entries, don't bump the busy count. Return a boolean indicating whether or not the busy count was bumped. (4) Break mac_policy_list into two lists: one with the same name holding dynamic policies, and a new list, mac_static_policy_list, which holds policies loaded before mac_late and without the unload flag set. The static list may be accessed without holding the busy count, since it can't change at run-time. (5) In general, prefer making the list busy conditionally, meaning we pay only one mutex lock per entry point if all modules are on the static list, rather than two (since we don't have to lower the busy count when we're done with the framework). For systems running just Biba or MLS, this will halve the mutex accesses in the network stack, and may offer a substantial performance benefits. (6) Lay the groundwork for a dynamic-free kernel option which eliminates all locking associated with dynamically loaded or unloaded policies, for pre-configured systems requiring maximum performance but less run-time flexibility. These changes have been running for a few weeks on MAC development branch systems. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-07 17:49:24 +00:00
int entrycount; \
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
\
Clean up locking for the MAC Framework: (1) Accept that we're now going to use mutexes, so don't attempt to avoid treating them as mutexes. This cleans up locking accessor function names some. (2) Rename variables to _mtx, _cv, _count, simplifying the naming. (3) Add a new form of the _busy() primitive that conditionally makes the list busy: if there are entries on the list, bump the busy count. If there are no entries, don't bump the busy count. Return a boolean indicating whether or not the busy count was bumped. (4) Break mac_policy_list into two lists: one with the same name holding dynamic policies, and a new list, mac_static_policy_list, which holds policies loaded before mac_late and without the unload flag set. The static list may be accessed without holding the busy count, since it can't change at run-time. (5) In general, prefer making the list busy conditionally, meaning we pay only one mutex lock per entry point if all modules are on the static list, rather than two (since we don't have to lower the busy count when we're done with the framework). For systems running just Biba or MLS, this will halve the mutex accesses in the network stack, and may offer a substantial performance benefits. (6) Lay the groundwork for a dynamic-free kernel option which eliminates all locking associated with dynamically loaded or unloaded policies, for pre-configured systems requiring maximum performance but less run-time flexibility. These changes have been running for a few weeks on MAC development branch systems. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-07 17:49:24 +00:00
LIST_FOREACH(mpc, &mac_static_policy_list, mpc_list) { \
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
if (mpc->mpc_ops->mpo_ ## operation != NULL) \
result = result composition \
mpc->mpc_ops->mpo_ ## operation (args); \
} \
Clean up locking for the MAC Framework: (1) Accept that we're now going to use mutexes, so don't attempt to avoid treating them as mutexes. This cleans up locking accessor function names some. (2) Rename variables to _mtx, _cv, _count, simplifying the naming. (3) Add a new form of the _busy() primitive that conditionally makes the list busy: if there are entries on the list, bump the busy count. If there are no entries, don't bump the busy count. Return a boolean indicating whether or not the busy count was bumped. (4) Break mac_policy_list into two lists: one with the same name holding dynamic policies, and a new list, mac_static_policy_list, which holds policies loaded before mac_late and without the unload flag set. The static list may be accessed without holding the busy count, since it can't change at run-time. (5) In general, prefer making the list busy conditionally, meaning we pay only one mutex lock per entry point if all modules are on the static list, rather than two (since we don't have to lower the busy count when we're done with the framework). For systems running just Biba or MLS, this will halve the mutex accesses in the network stack, and may offer a substantial performance benefits. (6) Lay the groundwork for a dynamic-free kernel option which eliminates all locking associated with dynamically loaded or unloaded policies, for pre-configured systems requiring maximum performance but less run-time flexibility. These changes have been running for a few weeks on MAC development branch systems. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-07 17:49:24 +00:00
if ((entrycount = mac_policy_list_conditional_busy()) != 0) { \
LIST_FOREACH(mpc, &mac_policy_list, mpc_list) { \
if (mpc->mpc_ops->mpo_ ## operation != NULL) \
result = result composition \
mpc->mpc_ops->mpo_ ## operation \
(args); \
} \
mac_policy_list_unbusy(); \
} \
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
} while (0)
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
#define MAC_EXTERNALIZE(type, label, elementlist, outbuf, \
outbuflen) do { \
Redesign the externalization APIs from the MAC Framework to the MAC policy modules to improve robustness against C string bugs and vulnerabilities. Following these revisions, all string construction of labels for export to userspace (or elsewhere) is performed using the sbuf API, which prevents the consumer from having to perform laborious and intricate pointer and buffer checks. This substantially simplifies the externalization logic, both at the MAC Framework level, and in individual policies; this becomes especially useful when policies export more complex label data, such as with compartments in Biba and MLS. Bundled in here are some other minor fixes associated with externalization: including avoiding malloc while holding the process mutex in mac_lomac, and hence avoid a failure mode when printing labels during a downgrade operation due to the removal of the M_NOWAIT case. This has been running in the MAC development tree for about three weeks without problems. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-06-23 01:26:34 +00:00
int claimed, first, ignorenotfound, savedlen; \
char *element_name, *element_temp; \
struct sbuf sb; \
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
\
error = 0; \
first = 1; \
Redesign the externalization APIs from the MAC Framework to the MAC policy modules to improve robustness against C string bugs and vulnerabilities. Following these revisions, all string construction of labels for export to userspace (or elsewhere) is performed using the sbuf API, which prevents the consumer from having to perform laborious and intricate pointer and buffer checks. This substantially simplifies the externalization logic, both at the MAC Framework level, and in individual policies; this becomes especially useful when policies export more complex label data, such as with compartments in Biba and MLS. Bundled in here are some other minor fixes associated with externalization: including avoiding malloc while holding the process mutex in mac_lomac, and hence avoid a failure mode when printing labels during a downgrade operation due to the removal of the M_NOWAIT case. This has been running in the MAC development tree for about three weeks without problems. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-06-23 01:26:34 +00:00
sbuf_new(&sb, outbuf, outbuflen, SBUF_FIXEDLEN); \
element_temp = elementlist; \
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
while ((element_name = strsep(&element_temp, ",")) != NULL) { \
if (element_name[0] == '?') { \
element_name++; \
ignorenotfound = 1; \
Redesign the externalization APIs from the MAC Framework to the MAC policy modules to improve robustness against C string bugs and vulnerabilities. Following these revisions, all string construction of labels for export to userspace (or elsewhere) is performed using the sbuf API, which prevents the consumer from having to perform laborious and intricate pointer and buffer checks. This substantially simplifies the externalization logic, both at the MAC Framework level, and in individual policies; this becomes especially useful when policies export more complex label data, such as with compartments in Biba and MLS. Bundled in here are some other minor fixes associated with externalization: including avoiding malloc while holding the process mutex in mac_lomac, and hence avoid a failure mode when printing labels during a downgrade operation due to the removal of the M_NOWAIT case. This has been running in the MAC development tree for about three weeks without problems. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-06-23 01:26:34 +00:00
} else \
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
ignorenotfound = 0; \
Redesign the externalization APIs from the MAC Framework to the MAC policy modules to improve robustness against C string bugs and vulnerabilities. Following these revisions, all string construction of labels for export to userspace (or elsewhere) is performed using the sbuf API, which prevents the consumer from having to perform laborious and intricate pointer and buffer checks. This substantially simplifies the externalization logic, both at the MAC Framework level, and in individual policies; this becomes especially useful when policies export more complex label data, such as with compartments in Biba and MLS. Bundled in here are some other minor fixes associated with externalization: including avoiding malloc while holding the process mutex in mac_lomac, and hence avoid a failure mode when printing labels during a downgrade operation due to the removal of the M_NOWAIT case. This has been running in the MAC development tree for about three weeks without problems. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-06-23 01:26:34 +00:00
savedlen = sbuf_len(&sb); \
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
if (first) { \
Redesign the externalization APIs from the MAC Framework to the MAC policy modules to improve robustness against C string bugs and vulnerabilities. Following these revisions, all string construction of labels for export to userspace (or elsewhere) is performed using the sbuf API, which prevents the consumer from having to perform laborious and intricate pointer and buffer checks. This substantially simplifies the externalization logic, both at the MAC Framework level, and in individual policies; this becomes especially useful when policies export more complex label data, such as with compartments in Biba and MLS. Bundled in here are some other minor fixes associated with externalization: including avoiding malloc while holding the process mutex in mac_lomac, and hence avoid a failure mode when printing labels during a downgrade operation due to the removal of the M_NOWAIT case. This has been running in the MAC development tree for about three weeks without problems. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-06-23 01:26:34 +00:00
error = sbuf_printf(&sb, "%s/", element_name); \
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
first = 0; \
} else \
Redesign the externalization APIs from the MAC Framework to the MAC policy modules to improve robustness against C string bugs and vulnerabilities. Following these revisions, all string construction of labels for export to userspace (or elsewhere) is performed using the sbuf API, which prevents the consumer from having to perform laborious and intricate pointer and buffer checks. This substantially simplifies the externalization logic, both at the MAC Framework level, and in individual policies; this becomes especially useful when policies export more complex label data, such as with compartments in Biba and MLS. Bundled in here are some other minor fixes associated with externalization: including avoiding malloc while holding the process mutex in mac_lomac, and hence avoid a failure mode when printing labels during a downgrade operation due to the removal of the M_NOWAIT case. This has been running in the MAC development tree for about three weeks without problems. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-06-23 01:26:34 +00:00
error = sbuf_printf(&sb, ",%s/", element_name); \
if (error == -1) { \
error = EINVAL; /* XXX: E2BIG? */ \
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
break; \
} \
Redesign the externalization APIs from the MAC Framework to the MAC policy modules to improve robustness against C string bugs and vulnerabilities. Following these revisions, all string construction of labels for export to userspace (or elsewhere) is performed using the sbuf API, which prevents the consumer from having to perform laborious and intricate pointer and buffer checks. This substantially simplifies the externalization logic, both at the MAC Framework level, and in individual policies; this becomes especially useful when policies export more complex label data, such as with compartments in Biba and MLS. Bundled in here are some other minor fixes associated with externalization: including avoiding malloc while holding the process mutex in mac_lomac, and hence avoid a failure mode when printing labels during a downgrade operation due to the removal of the M_NOWAIT case. This has been running in the MAC development tree for about three weeks without problems. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-06-23 01:26:34 +00:00
claimed = 0; \
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
MAC_CHECK(externalize_ ## type, label, element_name, \
Redesign the externalization APIs from the MAC Framework to the MAC policy modules to improve robustness against C string bugs and vulnerabilities. Following these revisions, all string construction of labels for export to userspace (or elsewhere) is performed using the sbuf API, which prevents the consumer from having to perform laborious and intricate pointer and buffer checks. This substantially simplifies the externalization logic, both at the MAC Framework level, and in individual policies; this becomes especially useful when policies export more complex label data, such as with compartments in Biba and MLS. Bundled in here are some other minor fixes associated with externalization: including avoiding malloc while holding the process mutex in mac_lomac, and hence avoid a failure mode when printing labels during a downgrade operation due to the removal of the M_NOWAIT case. This has been running in the MAC development tree for about three weeks without problems. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-06-23 01:26:34 +00:00
&sb, &claimed); \
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
if (error) \
break; \
Redesign the externalization APIs from the MAC Framework to the MAC policy modules to improve robustness against C string bugs and vulnerabilities. Following these revisions, all string construction of labels for export to userspace (or elsewhere) is performed using the sbuf API, which prevents the consumer from having to perform laborious and intricate pointer and buffer checks. This substantially simplifies the externalization logic, both at the MAC Framework level, and in individual policies; this becomes especially useful when policies export more complex label data, such as with compartments in Biba and MLS. Bundled in here are some other minor fixes associated with externalization: including avoiding malloc while holding the process mutex in mac_lomac, and hence avoid a failure mode when printing labels during a downgrade operation due to the removal of the M_NOWAIT case. This has been running in the MAC development tree for about three weeks without problems. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-06-23 01:26:34 +00:00
if (claimed == 0 && ignorenotfound) { \
/* Revert last label name. */ \
sbuf_setpos(&sb, savedlen); \
} else if (claimed != 1) { \
error = EINVAL; /* XXX: ENOLABEL? */ \
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
break; \
} \
} \
Redesign the externalization APIs from the MAC Framework to the MAC policy modules to improve robustness against C string bugs and vulnerabilities. Following these revisions, all string construction of labels for export to userspace (or elsewhere) is performed using the sbuf API, which prevents the consumer from having to perform laborious and intricate pointer and buffer checks. This substantially simplifies the externalization logic, both at the MAC Framework level, and in individual policies; this becomes especially useful when policies export more complex label data, such as with compartments in Biba and MLS. Bundled in here are some other minor fixes associated with externalization: including avoiding malloc while holding the process mutex in mac_lomac, and hence avoid a failure mode when printing labels during a downgrade operation due to the removal of the M_NOWAIT case. This has been running in the MAC development tree for about three weeks without problems. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-06-23 01:26:34 +00:00
sbuf_finish(&sb); \
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
} while (0)
#define MAC_INTERNALIZE(type, label, instring) do { \
char *element, *element_name, *element_data; \
int claimed; \
\
error = 0; \
element = instring; \
while ((element_name = strsep(&element, ",")) != NULL) { \
element_data = element_name; \
element_name = strsep(&element_data, "/"); \
if (element_data == NULL) { \
error = EINVAL; \
break; \
} \
claimed = 0; \
MAC_CHECK(internalize_ ## type, label, element_name, \
element_data, &claimed); \
if (error) \
break; \
if (claimed != 1) { \
/* XXXMAC: Another error here? */ \
error = EINVAL; \
break; \
} \
} \
} while (0)
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
/*
* MAC_PERFORM performs the designated operation by walking the policy
* module list and invoking that operation for each policy.
*/
#define MAC_PERFORM(operation, args...) do { \
struct mac_policy_conf *mpc; \
Clean up locking for the MAC Framework: (1) Accept that we're now going to use mutexes, so don't attempt to avoid treating them as mutexes. This cleans up locking accessor function names some. (2) Rename variables to _mtx, _cv, _count, simplifying the naming. (3) Add a new form of the _busy() primitive that conditionally makes the list busy: if there are entries on the list, bump the busy count. If there are no entries, don't bump the busy count. Return a boolean indicating whether or not the busy count was bumped. (4) Break mac_policy_list into two lists: one with the same name holding dynamic policies, and a new list, mac_static_policy_list, which holds policies loaded before mac_late and without the unload flag set. The static list may be accessed without holding the busy count, since it can't change at run-time. (5) In general, prefer making the list busy conditionally, meaning we pay only one mutex lock per entry point if all modules are on the static list, rather than two (since we don't have to lower the busy count when we're done with the framework). For systems running just Biba or MLS, this will halve the mutex accesses in the network stack, and may offer a substantial performance benefits. (6) Lay the groundwork for a dynamic-free kernel option which eliminates all locking associated with dynamically loaded or unloaded policies, for pre-configured systems requiring maximum performance but less run-time flexibility. These changes have been running for a few weeks on MAC development branch systems. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-07 17:49:24 +00:00
int entrycount; \
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
\
Clean up locking for the MAC Framework: (1) Accept that we're now going to use mutexes, so don't attempt to avoid treating them as mutexes. This cleans up locking accessor function names some. (2) Rename variables to _mtx, _cv, _count, simplifying the naming. (3) Add a new form of the _busy() primitive that conditionally makes the list busy: if there are entries on the list, bump the busy count. If there are no entries, don't bump the busy count. Return a boolean indicating whether or not the busy count was bumped. (4) Break mac_policy_list into two lists: one with the same name holding dynamic policies, and a new list, mac_static_policy_list, which holds policies loaded before mac_late and without the unload flag set. The static list may be accessed without holding the busy count, since it can't change at run-time. (5) In general, prefer making the list busy conditionally, meaning we pay only one mutex lock per entry point if all modules are on the static list, rather than two (since we don't have to lower the busy count when we're done with the framework). For systems running just Biba or MLS, this will halve the mutex accesses in the network stack, and may offer a substantial performance benefits. (6) Lay the groundwork for a dynamic-free kernel option which eliminates all locking associated with dynamically loaded or unloaded policies, for pre-configured systems requiring maximum performance but less run-time flexibility. These changes have been running for a few weeks on MAC development branch systems. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-07 17:49:24 +00:00
LIST_FOREACH(mpc, &mac_static_policy_list, mpc_list) { \
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
if (mpc->mpc_ops->mpo_ ## operation != NULL) \
mpc->mpc_ops->mpo_ ## operation (args); \
} \
Clean up locking for the MAC Framework: (1) Accept that we're now going to use mutexes, so don't attempt to avoid treating them as mutexes. This cleans up locking accessor function names some. (2) Rename variables to _mtx, _cv, _count, simplifying the naming. (3) Add a new form of the _busy() primitive that conditionally makes the list busy: if there are entries on the list, bump the busy count. If there are no entries, don't bump the busy count. Return a boolean indicating whether or not the busy count was bumped. (4) Break mac_policy_list into two lists: one with the same name holding dynamic policies, and a new list, mac_static_policy_list, which holds policies loaded before mac_late and without the unload flag set. The static list may be accessed without holding the busy count, since it can't change at run-time. (5) In general, prefer making the list busy conditionally, meaning we pay only one mutex lock per entry point if all modules are on the static list, rather than two (since we don't have to lower the busy count when we're done with the framework). For systems running just Biba or MLS, this will halve the mutex accesses in the network stack, and may offer a substantial performance benefits. (6) Lay the groundwork for a dynamic-free kernel option which eliminates all locking associated with dynamically loaded or unloaded policies, for pre-configured systems requiring maximum performance but less run-time flexibility. These changes have been running for a few weeks on MAC development branch systems. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-07 17:49:24 +00:00
if ((entrycount = mac_policy_list_conditional_busy()) != 0) { \
LIST_FOREACH(mpc, &mac_policy_list, mpc_list) { \
if (mpc->mpc_ops->mpo_ ## operation != NULL) \
mpc->mpc_ops->mpo_ ## operation (args); \
} \
mac_policy_list_unbusy(); \
} \
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
} while (0)
/*
* Initialize the MAC subsystem, including appropriate SMP locks.
*/
static void
mac_init(void)
{
Clean up locking for the MAC Framework: (1) Accept that we're now going to use mutexes, so don't attempt to avoid treating them as mutexes. This cleans up locking accessor function names some. (2) Rename variables to _mtx, _cv, _count, simplifying the naming. (3) Add a new form of the _busy() primitive that conditionally makes the list busy: if there are entries on the list, bump the busy count. If there are no entries, don't bump the busy count. Return a boolean indicating whether or not the busy count was bumped. (4) Break mac_policy_list into two lists: one with the same name holding dynamic policies, and a new list, mac_static_policy_list, which holds policies loaded before mac_late and without the unload flag set. The static list may be accessed without holding the busy count, since it can't change at run-time. (5) In general, prefer making the list busy conditionally, meaning we pay only one mutex lock per entry point if all modules are on the static list, rather than two (since we don't have to lower the busy count when we're done with the framework). For systems running just Biba or MLS, this will halve the mutex accesses in the network stack, and may offer a substantial performance benefits. (6) Lay the groundwork for a dynamic-free kernel option which eliminates all locking associated with dynamically loaded or unloaded policies, for pre-configured systems requiring maximum performance but less run-time flexibility. These changes have been running for a few weeks on MAC development branch systems. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-07 17:49:24 +00:00
LIST_INIT(&mac_static_policy_list);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
LIST_INIT(&mac_policy_list);
Clean up locking for the MAC Framework: (1) Accept that we're now going to use mutexes, so don't attempt to avoid treating them as mutexes. This cleans up locking accessor function names some. (2) Rename variables to _mtx, _cv, _count, simplifying the naming. (3) Add a new form of the _busy() primitive that conditionally makes the list busy: if there are entries on the list, bump the busy count. If there are no entries, don't bump the busy count. Return a boolean indicating whether or not the busy count was bumped. (4) Break mac_policy_list into two lists: one with the same name holding dynamic policies, and a new list, mac_static_policy_list, which holds policies loaded before mac_late and without the unload flag set. The static list may be accessed without holding the busy count, since it can't change at run-time. (5) In general, prefer making the list busy conditionally, meaning we pay only one mutex lock per entry point if all modules are on the static list, rather than two (since we don't have to lower the busy count when we're done with the framework). For systems running just Biba or MLS, this will halve the mutex accesses in the network stack, and may offer a substantial performance benefits. (6) Lay the groundwork for a dynamic-free kernel option which eliminates all locking associated with dynamically loaded or unloaded policies, for pre-configured systems requiring maximum performance but less run-time flexibility. These changes have been running for a few weeks on MAC development branch systems. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-07 17:49:24 +00:00
mtx_init(&mac_policy_mtx, "mac_policy_mtx", NULL, MTX_DEF);
cv_init(&mac_policy_cv, "mac_policy_cv");
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
/*
* For the purposes of modules that want to know if they were loaded
* "early", set the mac_late flag once we've processed modules either
* linked into the kernel, or loaded before the kernel startup.
*/
static void
mac_late_init(void)
{
mac_late = 1;
}
Move MAC label storage for mbufs into m_tags from the m_pkthdr structure, returning some additional room in the first mbuf in a chain, and avoiding feature-specific contents in the mbuf header. To do this: - Modify mbuf_to_label() to extract the tag, returning NULL if not found. - Introduce mac_init_mbuf_tag() which does most of the work mac_init_mbuf() used to do, except on an m_tag rather than an mbuf. - Scale back mac_init_mbuf() to perform m_tag allocation and invoke mac_init_mbuf_tag(). - Replace mac_destroy_mbuf() with mac_destroy_mbuf_tag(), since m_tag's are now GC'd deep in the m_tag/mbuf code rather than at a higher level when mbufs are directly free()'d. - Add mac_copy_mbuf_tag() to support m_copy_pkthdr() and related notions. - Generally change all references to mbuf labels so that they use mbuf_to_label() rather than &mbuf->m_pkthdr.label. This required no changes in the MAC policies (yay!). - Tweak mbuf release routines to not call mac_destroy_mbuf(), tag destruction takes care of it for us now. - Remove MAC magic from m_copy_pkthdr() and m_move_pkthdr() -- the existing m_tag support does all this for us. Note that we can no longer just zero the m_tag list on the target mbuf, rather, we have to delete the chain because m_tag's will already be hung off freshly allocated mbuf's. - Tweak m_tag copying routines so that if we're copying a MAC m_tag, we don't do a binary copy, rather, we initialize the new storage and do a deep copy of the label. - Remove use of MAC_FLAG_INITIALIZED in a few bizarre places having to do with mbuf header copies previously. - When an mbuf is copied in ip_input(), we no longer need to explicitly copy the label because it will get handled by the m_tag code now. - No longer any weird handling of MAC labels in if_loop.c during header copies. - Add MPC_LOADTIME_FLAG_LABELMBUFS flag to Biba, MLS, mac_test. In mac_test, handle the label==NULL case, since it can be dynamically loaded. In order to improve performance with this change, introduce the notion of "lazy MAC label allocation" -- only allocate m_tag storage for MAC labels if we're running with a policy that uses MAC labels on mbufs. Policies declare this intent by setting the MPC_LOADTIME_FLAG_LABELMBUFS flag in their load-time flags field during declaration. Note: this opens up the possibility of post-boot policy modules getting back NULL slot entries even though they have policy invariants of non-NULL slot entries, as the policy might have been loaded after the mbuf was allocated, leaving the mbuf without label storage. Policies that cannot handle this case must be declared as NOTLATE, or must be modified. - mac_labelmbufs holds the current cumulative status as to whether any policies require mbuf labeling or not. This is updated whenever the active policy set changes by the function mac_policy_updateflags(). The function iterates the list and checks whether any have the flag set. Write access to this variable is protected by the policy list; read access is currently not protected for performance reasons. This might change if it causes problems. - Add MAC_POLICY_LIST_ASSERT_EXCLUSIVE() to permit the flags update function to assert appropriate locks. - This makes allocation in mac_init_mbuf() conditional on the flag. Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 20:39:06 +00:00
/*
* After the policy list has changed, walk the list to update any global
Attempt to simplify #ifdef logic for MAC_ALWAYS_LABEL_MBUF. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-08-01 15:45:14 +00:00
* flags. Currently, we support only one flag, and it's conditionally
* defined; as a result, the entire function is conditional. Eventually,
* the #else case might also iterate across the policies.
Move MAC label storage for mbufs into m_tags from the m_pkthdr structure, returning some additional room in the first mbuf in a chain, and avoiding feature-specific contents in the mbuf header. To do this: - Modify mbuf_to_label() to extract the tag, returning NULL if not found. - Introduce mac_init_mbuf_tag() which does most of the work mac_init_mbuf() used to do, except on an m_tag rather than an mbuf. - Scale back mac_init_mbuf() to perform m_tag allocation and invoke mac_init_mbuf_tag(). - Replace mac_destroy_mbuf() with mac_destroy_mbuf_tag(), since m_tag's are now GC'd deep in the m_tag/mbuf code rather than at a higher level when mbufs are directly free()'d. - Add mac_copy_mbuf_tag() to support m_copy_pkthdr() and related notions. - Generally change all references to mbuf labels so that they use mbuf_to_label() rather than &mbuf->m_pkthdr.label. This required no changes in the MAC policies (yay!). - Tweak mbuf release routines to not call mac_destroy_mbuf(), tag destruction takes care of it for us now. - Remove MAC magic from m_copy_pkthdr() and m_move_pkthdr() -- the existing m_tag support does all this for us. Note that we can no longer just zero the m_tag list on the target mbuf, rather, we have to delete the chain because m_tag's will already be hung off freshly allocated mbuf's. - Tweak m_tag copying routines so that if we're copying a MAC m_tag, we don't do a binary copy, rather, we initialize the new storage and do a deep copy of the label. - Remove use of MAC_FLAG_INITIALIZED in a few bizarre places having to do with mbuf header copies previously. - When an mbuf is copied in ip_input(), we no longer need to explicitly copy the label because it will get handled by the m_tag code now. - No longer any weird handling of MAC labels in if_loop.c during header copies. - Add MPC_LOADTIME_FLAG_LABELMBUFS flag to Biba, MLS, mac_test. In mac_test, handle the label==NULL case, since it can be dynamically loaded. In order to improve performance with this change, introduce the notion of "lazy MAC label allocation" -- only allocate m_tag storage for MAC labels if we're running with a policy that uses MAC labels on mbufs. Policies declare this intent by setting the MPC_LOADTIME_FLAG_LABELMBUFS flag in their load-time flags field during declaration. Note: this opens up the possibility of post-boot policy modules getting back NULL slot entries even though they have policy invariants of non-NULL slot entries, as the policy might have been loaded after the mbuf was allocated, leaving the mbuf without label storage. Policies that cannot handle this case must be declared as NOTLATE, or must be modified. - mac_labelmbufs holds the current cumulative status as to whether any policies require mbuf labeling or not. This is updated whenever the active policy set changes by the function mac_policy_updateflags(). The function iterates the list and checks whether any have the flag set. Write access to this variable is protected by the policy list; read access is currently not protected for performance reasons. This might change if it causes problems. - Add MAC_POLICY_LIST_ASSERT_EXCLUSIVE() to permit the flags update function to assert appropriate locks. - This makes allocation in mac_init_mbuf() conditional on the flag. Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 20:39:06 +00:00
*/
static void
mac_policy_updateflags(void)
{
#ifndef MAC_ALWAYS_LABEL_MBUF
Attempt to simplify #ifdef logic for MAC_ALWAYS_LABEL_MBUF. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-08-01 15:45:14 +00:00
struct mac_policy_conf *tmpc;
Move MAC label storage for mbufs into m_tags from the m_pkthdr structure, returning some additional room in the first mbuf in a chain, and avoiding feature-specific contents in the mbuf header. To do this: - Modify mbuf_to_label() to extract the tag, returning NULL if not found. - Introduce mac_init_mbuf_tag() which does most of the work mac_init_mbuf() used to do, except on an m_tag rather than an mbuf. - Scale back mac_init_mbuf() to perform m_tag allocation and invoke mac_init_mbuf_tag(). - Replace mac_destroy_mbuf() with mac_destroy_mbuf_tag(), since m_tag's are now GC'd deep in the m_tag/mbuf code rather than at a higher level when mbufs are directly free()'d. - Add mac_copy_mbuf_tag() to support m_copy_pkthdr() and related notions. - Generally change all references to mbuf labels so that they use mbuf_to_label() rather than &mbuf->m_pkthdr.label. This required no changes in the MAC policies (yay!). - Tweak mbuf release routines to not call mac_destroy_mbuf(), tag destruction takes care of it for us now. - Remove MAC magic from m_copy_pkthdr() and m_move_pkthdr() -- the existing m_tag support does all this for us. Note that we can no longer just zero the m_tag list on the target mbuf, rather, we have to delete the chain because m_tag's will already be hung off freshly allocated mbuf's. - Tweak m_tag copying routines so that if we're copying a MAC m_tag, we don't do a binary copy, rather, we initialize the new storage and do a deep copy of the label. - Remove use of MAC_FLAG_INITIALIZED in a few bizarre places having to do with mbuf header copies previously. - When an mbuf is copied in ip_input(), we no longer need to explicitly copy the label because it will get handled by the m_tag code now. - No longer any weird handling of MAC labels in if_loop.c during header copies. - Add MPC_LOADTIME_FLAG_LABELMBUFS flag to Biba, MLS, mac_test. In mac_test, handle the label==NULL case, since it can be dynamically loaded. In order to improve performance with this change, introduce the notion of "lazy MAC label allocation" -- only allocate m_tag storage for MAC labels if we're running with a policy that uses MAC labels on mbufs. Policies declare this intent by setting the MPC_LOADTIME_FLAG_LABELMBUFS flag in their load-time flags field during declaration. Note: this opens up the possibility of post-boot policy modules getting back NULL slot entries even though they have policy invariants of non-NULL slot entries, as the policy might have been loaded after the mbuf was allocated, leaving the mbuf without label storage. Policies that cannot handle this case must be declared as NOTLATE, or must be modified. - mac_labelmbufs holds the current cumulative status as to whether any policies require mbuf labeling or not. This is updated whenever the active policy set changes by the function mac_policy_updateflags(). The function iterates the list and checks whether any have the flag set. Write access to this variable is protected by the policy list; read access is currently not protected for performance reasons. This might change if it causes problems. - Add MAC_POLICY_LIST_ASSERT_EXCLUSIVE() to permit the flags update function to assert appropriate locks. - This makes allocation in mac_init_mbuf() conditional on the flag. Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 20:39:06 +00:00
int labelmbufs;
Clean up locking for the MAC Framework: (1) Accept that we're now going to use mutexes, so don't attempt to avoid treating them as mutexes. This cleans up locking accessor function names some. (2) Rename variables to _mtx, _cv, _count, simplifying the naming. (3) Add a new form of the _busy() primitive that conditionally makes the list busy: if there are entries on the list, bump the busy count. If there are no entries, don't bump the busy count. Return a boolean indicating whether or not the busy count was bumped. (4) Break mac_policy_list into two lists: one with the same name holding dynamic policies, and a new list, mac_static_policy_list, which holds policies loaded before mac_late and without the unload flag set. The static list may be accessed without holding the busy count, since it can't change at run-time. (5) In general, prefer making the list busy conditionally, meaning we pay only one mutex lock per entry point if all modules are on the static list, rather than two (since we don't have to lower the busy count when we're done with the framework). For systems running just Biba or MLS, this will halve the mutex accesses in the network stack, and may offer a substantial performance benefits. (6) Lay the groundwork for a dynamic-free kernel option which eliminates all locking associated with dynamically loaded or unloaded policies, for pre-configured systems requiring maximum performance but less run-time flexibility. These changes have been running for a few weeks on MAC development branch systems. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-07 17:49:24 +00:00
mac_policy_assert_exclusive();
Move MAC label storage for mbufs into m_tags from the m_pkthdr structure, returning some additional room in the first mbuf in a chain, and avoiding feature-specific contents in the mbuf header. To do this: - Modify mbuf_to_label() to extract the tag, returning NULL if not found. - Introduce mac_init_mbuf_tag() which does most of the work mac_init_mbuf() used to do, except on an m_tag rather than an mbuf. - Scale back mac_init_mbuf() to perform m_tag allocation and invoke mac_init_mbuf_tag(). - Replace mac_destroy_mbuf() with mac_destroy_mbuf_tag(), since m_tag's are now GC'd deep in the m_tag/mbuf code rather than at a higher level when mbufs are directly free()'d. - Add mac_copy_mbuf_tag() to support m_copy_pkthdr() and related notions. - Generally change all references to mbuf labels so that they use mbuf_to_label() rather than &mbuf->m_pkthdr.label. This required no changes in the MAC policies (yay!). - Tweak mbuf release routines to not call mac_destroy_mbuf(), tag destruction takes care of it for us now. - Remove MAC magic from m_copy_pkthdr() and m_move_pkthdr() -- the existing m_tag support does all this for us. Note that we can no longer just zero the m_tag list on the target mbuf, rather, we have to delete the chain because m_tag's will already be hung off freshly allocated mbuf's. - Tweak m_tag copying routines so that if we're copying a MAC m_tag, we don't do a binary copy, rather, we initialize the new storage and do a deep copy of the label. - Remove use of MAC_FLAG_INITIALIZED in a few bizarre places having to do with mbuf header copies previously. - When an mbuf is copied in ip_input(), we no longer need to explicitly copy the label because it will get handled by the m_tag code now. - No longer any weird handling of MAC labels in if_loop.c during header copies. - Add MPC_LOADTIME_FLAG_LABELMBUFS flag to Biba, MLS, mac_test. In mac_test, handle the label==NULL case, since it can be dynamically loaded. In order to improve performance with this change, introduce the notion of "lazy MAC label allocation" -- only allocate m_tag storage for MAC labels if we're running with a policy that uses MAC labels on mbufs. Policies declare this intent by setting the MPC_LOADTIME_FLAG_LABELMBUFS flag in their load-time flags field during declaration. Note: this opens up the possibility of post-boot policy modules getting back NULL slot entries even though they have policy invariants of non-NULL slot entries, as the policy might have been loaded after the mbuf was allocated, leaving the mbuf without label storage. Policies that cannot handle this case must be declared as NOTLATE, or must be modified. - mac_labelmbufs holds the current cumulative status as to whether any policies require mbuf labeling or not. This is updated whenever the active policy set changes by the function mac_policy_updateflags(). The function iterates the list and checks whether any have the flag set. Write access to this variable is protected by the policy list; read access is currently not protected for performance reasons. This might change if it causes problems. - Add MAC_POLICY_LIST_ASSERT_EXCLUSIVE() to permit the flags update function to assert appropriate locks. - This makes allocation in mac_init_mbuf() conditional on the flag. Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 20:39:06 +00:00
labelmbufs = 0;
Clean up locking for the MAC Framework: (1) Accept that we're now going to use mutexes, so don't attempt to avoid treating them as mutexes. This cleans up locking accessor function names some. (2) Rename variables to _mtx, _cv, _count, simplifying the naming. (3) Add a new form of the _busy() primitive that conditionally makes the list busy: if there are entries on the list, bump the busy count. If there are no entries, don't bump the busy count. Return a boolean indicating whether or not the busy count was bumped. (4) Break mac_policy_list into two lists: one with the same name holding dynamic policies, and a new list, mac_static_policy_list, which holds policies loaded before mac_late and without the unload flag set. The static list may be accessed without holding the busy count, since it can't change at run-time. (5) In general, prefer making the list busy conditionally, meaning we pay only one mutex lock per entry point if all modules are on the static list, rather than two (since we don't have to lower the busy count when we're done with the framework). For systems running just Biba or MLS, this will halve the mutex accesses in the network stack, and may offer a substantial performance benefits. (6) Lay the groundwork for a dynamic-free kernel option which eliminates all locking associated with dynamically loaded or unloaded policies, for pre-configured systems requiring maximum performance but less run-time flexibility. These changes have been running for a few weeks on MAC development branch systems. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-07 17:49:24 +00:00
LIST_FOREACH(tmpc, &mac_static_policy_list, mpc_list) {
if (tmpc->mpc_loadtime_flags & MPC_LOADTIME_FLAG_LABELMBUFS)
labelmbufs++;
}
Move MAC label storage for mbufs into m_tags from the m_pkthdr structure, returning some additional room in the first mbuf in a chain, and avoiding feature-specific contents in the mbuf header. To do this: - Modify mbuf_to_label() to extract the tag, returning NULL if not found. - Introduce mac_init_mbuf_tag() which does most of the work mac_init_mbuf() used to do, except on an m_tag rather than an mbuf. - Scale back mac_init_mbuf() to perform m_tag allocation and invoke mac_init_mbuf_tag(). - Replace mac_destroy_mbuf() with mac_destroy_mbuf_tag(), since m_tag's are now GC'd deep in the m_tag/mbuf code rather than at a higher level when mbufs are directly free()'d. - Add mac_copy_mbuf_tag() to support m_copy_pkthdr() and related notions. - Generally change all references to mbuf labels so that they use mbuf_to_label() rather than &mbuf->m_pkthdr.label. This required no changes in the MAC policies (yay!). - Tweak mbuf release routines to not call mac_destroy_mbuf(), tag destruction takes care of it for us now. - Remove MAC magic from m_copy_pkthdr() and m_move_pkthdr() -- the existing m_tag support does all this for us. Note that we can no longer just zero the m_tag list on the target mbuf, rather, we have to delete the chain because m_tag's will already be hung off freshly allocated mbuf's. - Tweak m_tag copying routines so that if we're copying a MAC m_tag, we don't do a binary copy, rather, we initialize the new storage and do a deep copy of the label. - Remove use of MAC_FLAG_INITIALIZED in a few bizarre places having to do with mbuf header copies previously. - When an mbuf is copied in ip_input(), we no longer need to explicitly copy the label because it will get handled by the m_tag code now. - No longer any weird handling of MAC labels in if_loop.c during header copies. - Add MPC_LOADTIME_FLAG_LABELMBUFS flag to Biba, MLS, mac_test. In mac_test, handle the label==NULL case, since it can be dynamically loaded. In order to improve performance with this change, introduce the notion of "lazy MAC label allocation" -- only allocate m_tag storage for MAC labels if we're running with a policy that uses MAC labels on mbufs. Policies declare this intent by setting the MPC_LOADTIME_FLAG_LABELMBUFS flag in their load-time flags field during declaration. Note: this opens up the possibility of post-boot policy modules getting back NULL slot entries even though they have policy invariants of non-NULL slot entries, as the policy might have been loaded after the mbuf was allocated, leaving the mbuf without label storage. Policies that cannot handle this case must be declared as NOTLATE, or must be modified. - mac_labelmbufs holds the current cumulative status as to whether any policies require mbuf labeling or not. This is updated whenever the active policy set changes by the function mac_policy_updateflags(). The function iterates the list and checks whether any have the flag set. Write access to this variable is protected by the policy list; read access is currently not protected for performance reasons. This might change if it causes problems. - Add MAC_POLICY_LIST_ASSERT_EXCLUSIVE() to permit the flags update function to assert appropriate locks. - This makes allocation in mac_init_mbuf() conditional on the flag. Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 20:39:06 +00:00
LIST_FOREACH(tmpc, &mac_policy_list, mpc_list) {
if (tmpc->mpc_loadtime_flags & MPC_LOADTIME_FLAG_LABELMBUFS)
labelmbufs++;
}
mac_labelmbufs = (labelmbufs != 0);
#endif
}
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
/*
* Allow MAC policy modules to register during boot, etc.
*/
int
mac_policy_modevent(module_t mod, int type, void *data)
{
struct mac_policy_conf *mpc;
int error;
error = 0;
mpc = (struct mac_policy_conf *) data;
switch (type) {
case MOD_LOAD:
if (mpc->mpc_loadtime_flags & MPC_LOADTIME_FLAG_NOTLATE &&
mac_late) {
printf("mac_policy_modevent: can't load %s policy "
"after booting\n", mpc->mpc_name);
error = EBUSY;
break;
}
error = mac_policy_register(mpc);
break;
case MOD_UNLOAD:
/* Don't unregister the module if it was never registered. */
if ((mpc->mpc_runtime_flags & MPC_RUNTIME_FLAG_REGISTERED)
!= 0)
error = mac_policy_unregister(mpc);
else
error = 0;
break;
default:
break;
}
return (error);
}
static int
mac_policy_register(struct mac_policy_conf *mpc)
{
struct mac_policy_conf *tmpc;
Clean up locking for the MAC Framework: (1) Accept that we're now going to use mutexes, so don't attempt to avoid treating them as mutexes. This cleans up locking accessor function names some. (2) Rename variables to _mtx, _cv, _count, simplifying the naming. (3) Add a new form of the _busy() primitive that conditionally makes the list busy: if there are entries on the list, bump the busy count. If there are no entries, don't bump the busy count. Return a boolean indicating whether or not the busy count was bumped. (4) Break mac_policy_list into two lists: one with the same name holding dynamic policies, and a new list, mac_static_policy_list, which holds policies loaded before mac_late and without the unload flag set. The static list may be accessed without holding the busy count, since it can't change at run-time. (5) In general, prefer making the list busy conditionally, meaning we pay only one mutex lock per entry point if all modules are on the static list, rather than two (since we don't have to lower the busy count when we're done with the framework). For systems running just Biba or MLS, this will halve the mutex accesses in the network stack, and may offer a substantial performance benefits. (6) Lay the groundwork for a dynamic-free kernel option which eliminates all locking associated with dynamically loaded or unloaded policies, for pre-configured systems requiring maximum performance but less run-time flexibility. These changes have been running for a few weeks on MAC development branch systems. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-07 17:49:24 +00:00
int error, slot, static_entry;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
Clean up locking for the MAC Framework: (1) Accept that we're now going to use mutexes, so don't attempt to avoid treating them as mutexes. This cleans up locking accessor function names some. (2) Rename variables to _mtx, _cv, _count, simplifying the naming. (3) Add a new form of the _busy() primitive that conditionally makes the list busy: if there are entries on the list, bump the busy count. If there are no entries, don't bump the busy count. Return a boolean indicating whether or not the busy count was bumped. (4) Break mac_policy_list into two lists: one with the same name holding dynamic policies, and a new list, mac_static_policy_list, which holds policies loaded before mac_late and without the unload flag set. The static list may be accessed without holding the busy count, since it can't change at run-time. (5) In general, prefer making the list busy conditionally, meaning we pay only one mutex lock per entry point if all modules are on the static list, rather than two (since we don't have to lower the busy count when we're done with the framework). For systems running just Biba or MLS, this will halve the mutex accesses in the network stack, and may offer a substantial performance benefits. (6) Lay the groundwork for a dynamic-free kernel option which eliminates all locking associated with dynamically loaded or unloaded policies, for pre-configured systems requiring maximum performance but less run-time flexibility. These changes have been running for a few weeks on MAC development branch systems. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-07 17:49:24 +00:00
error = 0;
/*
* We don't technically need exclusive access while !mac_late,
* but hold it for assertion consistency.
*/
mac_policy_grab_exclusive();
/*
* If the module can potentially be unloaded, or we're loading
* late, we have to stick it in the non-static list and pay
* an extra performance overhead. Otherwise, we can pay a
* light locking cost and stick it in the static list.
*/
static_entry = (!mac_late &&
!(mpc->mpc_loadtime_flags & MPC_LOADTIME_FLAG_UNLOADOK));
if (static_entry) {
LIST_FOREACH(tmpc, &mac_static_policy_list, mpc_list) {
if (strcmp(tmpc->mpc_name, mpc->mpc_name) == 0) {
error = EEXIST;
goto out;
}
}
} else {
LIST_FOREACH(tmpc, &mac_policy_list, mpc_list) {
if (strcmp(tmpc->mpc_name, mpc->mpc_name) == 0) {
error = EEXIST;
goto out;
}
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
}
if (mpc->mpc_field_off != NULL) {
Rename MAC_MAX_POLICIES to MAC_MAX_SLOTS, since the variables and constants in question refer to the number of label slots, not the maximum number of policies that may be loaded. This should reduce confusion regarding an element in the MAC sysctl MIB, as well as make it more clear what the affect of changing the compile-time constants is. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-08 19:49:42 +00:00
slot = ffs(mac_slot_offsets_free);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
if (slot == 0) {
Clean up locking for the MAC Framework: (1) Accept that we're now going to use mutexes, so don't attempt to avoid treating them as mutexes. This cleans up locking accessor function names some. (2) Rename variables to _mtx, _cv, _count, simplifying the naming. (3) Add a new form of the _busy() primitive that conditionally makes the list busy: if there are entries on the list, bump the busy count. If there are no entries, don't bump the busy count. Return a boolean indicating whether or not the busy count was bumped. (4) Break mac_policy_list into two lists: one with the same name holding dynamic policies, and a new list, mac_static_policy_list, which holds policies loaded before mac_late and without the unload flag set. The static list may be accessed without holding the busy count, since it can't change at run-time. (5) In general, prefer making the list busy conditionally, meaning we pay only one mutex lock per entry point if all modules are on the static list, rather than two (since we don't have to lower the busy count when we're done with the framework). For systems running just Biba or MLS, this will halve the mutex accesses in the network stack, and may offer a substantial performance benefits. (6) Lay the groundwork for a dynamic-free kernel option which eliminates all locking associated with dynamically loaded or unloaded policies, for pre-configured systems requiring maximum performance but less run-time flexibility. These changes have been running for a few weeks on MAC development branch systems. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-07 17:49:24 +00:00
error = ENOMEM;
goto out;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
slot--;
Rename MAC_MAX_POLICIES to MAC_MAX_SLOTS, since the variables and constants in question refer to the number of label slots, not the maximum number of policies that may be loaded. This should reduce confusion regarding an element in the MAC sysctl MIB, as well as make it more clear what the affect of changing the compile-time constants is. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-08 19:49:42 +00:00
mac_slot_offsets_free &= ~(1 << slot);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
*mpc->mpc_field_off = slot;
}
mpc->mpc_runtime_flags |= MPC_RUNTIME_FLAG_REGISTERED;
Clean up locking for the MAC Framework: (1) Accept that we're now going to use mutexes, so don't attempt to avoid treating them as mutexes. This cleans up locking accessor function names some. (2) Rename variables to _mtx, _cv, _count, simplifying the naming. (3) Add a new form of the _busy() primitive that conditionally makes the list busy: if there are entries on the list, bump the busy count. If there are no entries, don't bump the busy count. Return a boolean indicating whether or not the busy count was bumped. (4) Break mac_policy_list into two lists: one with the same name holding dynamic policies, and a new list, mac_static_policy_list, which holds policies loaded before mac_late and without the unload flag set. The static list may be accessed without holding the busy count, since it can't change at run-time. (5) In general, prefer making the list busy conditionally, meaning we pay only one mutex lock per entry point if all modules are on the static list, rather than two (since we don't have to lower the busy count when we're done with the framework). For systems running just Biba or MLS, this will halve the mutex accesses in the network stack, and may offer a substantial performance benefits. (6) Lay the groundwork for a dynamic-free kernel option which eliminates all locking associated with dynamically loaded or unloaded policies, for pre-configured systems requiring maximum performance but less run-time flexibility. These changes have been running for a few weeks on MAC development branch systems. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-07 17:49:24 +00:00
/*
* If we're loading a MAC module after the framework has
* initialized, it has to go into the dynamic list. If
* we're loading it before we've finished initializing,
* it can go into the static list with weaker locker
* requirements.
*/
if (static_entry)
LIST_INSERT_HEAD(&mac_static_policy_list, mpc, mpc_list);
else
LIST_INSERT_HEAD(&mac_policy_list, mpc, mpc_list);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
/* Per-policy initialization. */
if (mpc->mpc_ops->mpo_init != NULL)
(*(mpc->mpc_ops->mpo_init))(mpc);
Move MAC label storage for mbufs into m_tags from the m_pkthdr structure, returning some additional room in the first mbuf in a chain, and avoiding feature-specific contents in the mbuf header. To do this: - Modify mbuf_to_label() to extract the tag, returning NULL if not found. - Introduce mac_init_mbuf_tag() which does most of the work mac_init_mbuf() used to do, except on an m_tag rather than an mbuf. - Scale back mac_init_mbuf() to perform m_tag allocation and invoke mac_init_mbuf_tag(). - Replace mac_destroy_mbuf() with mac_destroy_mbuf_tag(), since m_tag's are now GC'd deep in the m_tag/mbuf code rather than at a higher level when mbufs are directly free()'d. - Add mac_copy_mbuf_tag() to support m_copy_pkthdr() and related notions. - Generally change all references to mbuf labels so that they use mbuf_to_label() rather than &mbuf->m_pkthdr.label. This required no changes in the MAC policies (yay!). - Tweak mbuf release routines to not call mac_destroy_mbuf(), tag destruction takes care of it for us now. - Remove MAC magic from m_copy_pkthdr() and m_move_pkthdr() -- the existing m_tag support does all this for us. Note that we can no longer just zero the m_tag list on the target mbuf, rather, we have to delete the chain because m_tag's will already be hung off freshly allocated mbuf's. - Tweak m_tag copying routines so that if we're copying a MAC m_tag, we don't do a binary copy, rather, we initialize the new storage and do a deep copy of the label. - Remove use of MAC_FLAG_INITIALIZED in a few bizarre places having to do with mbuf header copies previously. - When an mbuf is copied in ip_input(), we no longer need to explicitly copy the label because it will get handled by the m_tag code now. - No longer any weird handling of MAC labels in if_loop.c during header copies. - Add MPC_LOADTIME_FLAG_LABELMBUFS flag to Biba, MLS, mac_test. In mac_test, handle the label==NULL case, since it can be dynamically loaded. In order to improve performance with this change, introduce the notion of "lazy MAC label allocation" -- only allocate m_tag storage for MAC labels if we're running with a policy that uses MAC labels on mbufs. Policies declare this intent by setting the MPC_LOADTIME_FLAG_LABELMBUFS flag in their load-time flags field during declaration. Note: this opens up the possibility of post-boot policy modules getting back NULL slot entries even though they have policy invariants of non-NULL slot entries, as the policy might have been loaded after the mbuf was allocated, leaving the mbuf without label storage. Policies that cannot handle this case must be declared as NOTLATE, or must be modified. - mac_labelmbufs holds the current cumulative status as to whether any policies require mbuf labeling or not. This is updated whenever the active policy set changes by the function mac_policy_updateflags(). The function iterates the list and checks whether any have the flag set. Write access to this variable is protected by the policy list; read access is currently not protected for performance reasons. This might change if it causes problems. - Add MAC_POLICY_LIST_ASSERT_EXCLUSIVE() to permit the flags update function to assert appropriate locks. - This makes allocation in mac_init_mbuf() conditional on the flag. Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 20:39:06 +00:00
mac_policy_updateflags();
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
printf("Security policy loaded: %s (%s)\n", mpc->mpc_fullname,
mpc->mpc_name);
Clean up locking for the MAC Framework: (1) Accept that we're now going to use mutexes, so don't attempt to avoid treating them as mutexes. This cleans up locking accessor function names some. (2) Rename variables to _mtx, _cv, _count, simplifying the naming. (3) Add a new form of the _busy() primitive that conditionally makes the list busy: if there are entries on the list, bump the busy count. If there are no entries, don't bump the busy count. Return a boolean indicating whether or not the busy count was bumped. (4) Break mac_policy_list into two lists: one with the same name holding dynamic policies, and a new list, mac_static_policy_list, which holds policies loaded before mac_late and without the unload flag set. The static list may be accessed without holding the busy count, since it can't change at run-time. (5) In general, prefer making the list busy conditionally, meaning we pay only one mutex lock per entry point if all modules are on the static list, rather than two (since we don't have to lower the busy count when we're done with the framework). For systems running just Biba or MLS, this will halve the mutex accesses in the network stack, and may offer a substantial performance benefits. (6) Lay the groundwork for a dynamic-free kernel option which eliminates all locking associated with dynamically loaded or unloaded policies, for pre-configured systems requiring maximum performance but less run-time flexibility. These changes have been running for a few weeks on MAC development branch systems. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-07 17:49:24 +00:00
out:
mac_policy_release_exclusive();
return (error);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
static int
mac_policy_unregister(struct mac_policy_conf *mpc)
{
Synch from TrustedBSD MAC tree: - If a policy isn't registered when a policy module unloads, silently succeed. - Hold the policy list lock across more of the validity tests to avoid races. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:46:03 +00:00
/*
* If we fail the load, we may get a request to unload. Check
* to see if we did the run-time registration, and if not,
* silently succeed.
*/
Clean up locking for the MAC Framework: (1) Accept that we're now going to use mutexes, so don't attempt to avoid treating them as mutexes. This cleans up locking accessor function names some. (2) Rename variables to _mtx, _cv, _count, simplifying the naming. (3) Add a new form of the _busy() primitive that conditionally makes the list busy: if there are entries on the list, bump the busy count. If there are no entries, don't bump the busy count. Return a boolean indicating whether or not the busy count was bumped. (4) Break mac_policy_list into two lists: one with the same name holding dynamic policies, and a new list, mac_static_policy_list, which holds policies loaded before mac_late and without the unload flag set. The static list may be accessed without holding the busy count, since it can't change at run-time. (5) In general, prefer making the list busy conditionally, meaning we pay only one mutex lock per entry point if all modules are on the static list, rather than two (since we don't have to lower the busy count when we're done with the framework). For systems running just Biba or MLS, this will halve the mutex accesses in the network stack, and may offer a substantial performance benefits. (6) Lay the groundwork for a dynamic-free kernel option which eliminates all locking associated with dynamically loaded or unloaded policies, for pre-configured systems requiring maximum performance but less run-time flexibility. These changes have been running for a few weeks on MAC development branch systems. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-07 17:49:24 +00:00
mac_policy_grab_exclusive();
Synch from TrustedBSD MAC tree: - If a policy isn't registered when a policy module unloads, silently succeed. - Hold the policy list lock across more of the validity tests to avoid races. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:46:03 +00:00
if ((mpc->mpc_runtime_flags & MPC_RUNTIME_FLAG_REGISTERED) == 0) {
Clean up locking for the MAC Framework: (1) Accept that we're now going to use mutexes, so don't attempt to avoid treating them as mutexes. This cleans up locking accessor function names some. (2) Rename variables to _mtx, _cv, _count, simplifying the naming. (3) Add a new form of the _busy() primitive that conditionally makes the list busy: if there are entries on the list, bump the busy count. If there are no entries, don't bump the busy count. Return a boolean indicating whether or not the busy count was bumped. (4) Break mac_policy_list into two lists: one with the same name holding dynamic policies, and a new list, mac_static_policy_list, which holds policies loaded before mac_late and without the unload flag set. The static list may be accessed without holding the busy count, since it can't change at run-time. (5) In general, prefer making the list busy conditionally, meaning we pay only one mutex lock per entry point if all modules are on the static list, rather than two (since we don't have to lower the busy count when we're done with the framework). For systems running just Biba or MLS, this will halve the mutex accesses in the network stack, and may offer a substantial performance benefits. (6) Lay the groundwork for a dynamic-free kernel option which eliminates all locking associated with dynamically loaded or unloaded policies, for pre-configured systems requiring maximum performance but less run-time flexibility. These changes have been running for a few weeks on MAC development branch systems. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-07 17:49:24 +00:00
mac_policy_release_exclusive();
Synch from TrustedBSD MAC tree: - If a policy isn't registered when a policy module unloads, silently succeed. - Hold the policy list lock across more of the validity tests to avoid races. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:46:03 +00:00
return (0);
}
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
#if 0
/*
* Don't allow unloading modules with private data.
*/
Synch from TrustedBSD MAC tree: - If a policy isn't registered when a policy module unloads, silently succeed. - Hold the policy list lock across more of the validity tests to avoid races. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:46:03 +00:00
if (mpc->mpc_field_off != NULL) {
MAC_POLICY_LIST_UNLOCK();
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
return (EBUSY);
Synch from TrustedBSD MAC tree: - If a policy isn't registered when a policy module unloads, silently succeed. - Hold the policy list lock across more of the validity tests to avoid races. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:46:03 +00:00
}
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
#endif
Synch from TrustedBSD MAC tree: - If a policy isn't registered when a policy module unloads, silently succeed. - Hold the policy list lock across more of the validity tests to avoid races. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:46:03 +00:00
/*
* Only allow the unload to proceed if the module is unloadable
* by its own definition.
*/
if ((mpc->mpc_loadtime_flags & MPC_LOADTIME_FLAG_UNLOADOK) == 0) {
Clean up locking for the MAC Framework: (1) Accept that we're now going to use mutexes, so don't attempt to avoid treating them as mutexes. This cleans up locking accessor function names some. (2) Rename variables to _mtx, _cv, _count, simplifying the naming. (3) Add a new form of the _busy() primitive that conditionally makes the list busy: if there are entries on the list, bump the busy count. If there are no entries, don't bump the busy count. Return a boolean indicating whether or not the busy count was bumped. (4) Break mac_policy_list into two lists: one with the same name holding dynamic policies, and a new list, mac_static_policy_list, which holds policies loaded before mac_late and without the unload flag set. The static list may be accessed without holding the busy count, since it can't change at run-time. (5) In general, prefer making the list busy conditionally, meaning we pay only one mutex lock per entry point if all modules are on the static list, rather than two (since we don't have to lower the busy count when we're done with the framework). For systems running just Biba or MLS, this will halve the mutex accesses in the network stack, and may offer a substantial performance benefits. (6) Lay the groundwork for a dynamic-free kernel option which eliminates all locking associated with dynamically loaded or unloaded policies, for pre-configured systems requiring maximum performance but less run-time flexibility. These changes have been running for a few weeks on MAC development branch systems. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-07 17:49:24 +00:00
mac_policy_release_exclusive();
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
return (EBUSY);
Synch from TrustedBSD MAC tree: - If a policy isn't registered when a policy module unloads, silently succeed. - Hold the policy list lock across more of the validity tests to avoid races. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:46:03 +00:00
}
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
if (mpc->mpc_ops->mpo_destroy != NULL)
(*(mpc->mpc_ops->mpo_destroy))(mpc);
LIST_REMOVE(mpc, mpc_list);
Make sure to clear the 'registered' flag for MAC policies when they unregister. Under some obscure (perhaps demented) circumstances, this can result in a panic if a policy is unregistered, and then someone foolishly unregisters it again. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-19 20:30:12 +00:00
mpc->mpc_runtime_flags &= ~MPC_RUNTIME_FLAG_REGISTERED;
Move MAC label storage for mbufs into m_tags from the m_pkthdr structure, returning some additional room in the first mbuf in a chain, and avoiding feature-specific contents in the mbuf header. To do this: - Modify mbuf_to_label() to extract the tag, returning NULL if not found. - Introduce mac_init_mbuf_tag() which does most of the work mac_init_mbuf() used to do, except on an m_tag rather than an mbuf. - Scale back mac_init_mbuf() to perform m_tag allocation and invoke mac_init_mbuf_tag(). - Replace mac_destroy_mbuf() with mac_destroy_mbuf_tag(), since m_tag's are now GC'd deep in the m_tag/mbuf code rather than at a higher level when mbufs are directly free()'d. - Add mac_copy_mbuf_tag() to support m_copy_pkthdr() and related notions. - Generally change all references to mbuf labels so that they use mbuf_to_label() rather than &mbuf->m_pkthdr.label. This required no changes in the MAC policies (yay!). - Tweak mbuf release routines to not call mac_destroy_mbuf(), tag destruction takes care of it for us now. - Remove MAC magic from m_copy_pkthdr() and m_move_pkthdr() -- the existing m_tag support does all this for us. Note that we can no longer just zero the m_tag list on the target mbuf, rather, we have to delete the chain because m_tag's will already be hung off freshly allocated mbuf's. - Tweak m_tag copying routines so that if we're copying a MAC m_tag, we don't do a binary copy, rather, we initialize the new storage and do a deep copy of the label. - Remove use of MAC_FLAG_INITIALIZED in a few bizarre places having to do with mbuf header copies previously. - When an mbuf is copied in ip_input(), we no longer need to explicitly copy the label because it will get handled by the m_tag code now. - No longer any weird handling of MAC labels in if_loop.c during header copies. - Add MPC_LOADTIME_FLAG_LABELMBUFS flag to Biba, MLS, mac_test. In mac_test, handle the label==NULL case, since it can be dynamically loaded. In order to improve performance with this change, introduce the notion of "lazy MAC label allocation" -- only allocate m_tag storage for MAC labels if we're running with a policy that uses MAC labels on mbufs. Policies declare this intent by setting the MPC_LOADTIME_FLAG_LABELMBUFS flag in their load-time flags field during declaration. Note: this opens up the possibility of post-boot policy modules getting back NULL slot entries even though they have policy invariants of non-NULL slot entries, as the policy might have been loaded after the mbuf was allocated, leaving the mbuf without label storage. Policies that cannot handle this case must be declared as NOTLATE, or must be modified. - mac_labelmbufs holds the current cumulative status as to whether any policies require mbuf labeling or not. This is updated whenever the active policy set changes by the function mac_policy_updateflags(). The function iterates the list and checks whether any have the flag set. Write access to this variable is protected by the policy list; read access is currently not protected for performance reasons. This might change if it causes problems. - Add MAC_POLICY_LIST_ASSERT_EXCLUSIVE() to permit the flags update function to assert appropriate locks. - This makes allocation in mac_init_mbuf() conditional on the flag. Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 20:39:06 +00:00
mac_policy_updateflags();
Clean up locking for the MAC Framework: (1) Accept that we're now going to use mutexes, so don't attempt to avoid treating them as mutexes. This cleans up locking accessor function names some. (2) Rename variables to _mtx, _cv, _count, simplifying the naming. (3) Add a new form of the _busy() primitive that conditionally makes the list busy: if there are entries on the list, bump the busy count. If there are no entries, don't bump the busy count. Return a boolean indicating whether or not the busy count was bumped. (4) Break mac_policy_list into two lists: one with the same name holding dynamic policies, and a new list, mac_static_policy_list, which holds policies loaded before mac_late and without the unload flag set. The static list may be accessed without holding the busy count, since it can't change at run-time. (5) In general, prefer making the list busy conditionally, meaning we pay only one mutex lock per entry point if all modules are on the static list, rather than two (since we don't have to lower the busy count when we're done with the framework). For systems running just Biba or MLS, this will halve the mutex accesses in the network stack, and may offer a substantial performance benefits. (6) Lay the groundwork for a dynamic-free kernel option which eliminates all locking associated with dynamically loaded or unloaded policies, for pre-configured systems requiring maximum performance but less run-time flexibility. These changes have been running for a few weeks on MAC development branch systems. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-07 17:49:24 +00:00
mac_policy_release_exclusive();
Introduce a condition variable to avoid returning EBUSY when the MAC policy list is busy during a load or unload attempt. We assert no locks held during the cv wait, meaning we should be fairly deadlock-safe. Because of the cv model and busy count, it's possible for a cv waiter waiting for exclusive access to the policy list to be starved by active and long-lived access control/labeling events. For now, we accept that as a necessary tradeoff. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-11-13 15:47:09 +00:00
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
printf("Security policy unload: %s (%s)\n", mpc->mpc_fullname,
mpc->mpc_name);
return (0);
}
/*
* Define an error value precedence, and given two arguments, selects the
* value with the higher precedence.
*/
static int
error_select(int error1, int error2)
{
/* Certain decision-making errors take top priority. */
if (error1 == EDEADLK || error2 == EDEADLK)
return (EDEADLK);
/* Invalid arguments should be reported where possible. */
if (error1 == EINVAL || error2 == EINVAL)
return (EINVAL);
/* Precedence goes to "visibility", with both process and file. */
if (error1 == ESRCH || error2 == ESRCH)
return (ESRCH);
if (error1 == ENOENT || error2 == ENOENT)
return (ENOENT);
/* Precedence goes to DAC/MAC protections. */
if (error1 == EACCES || error2 == EACCES)
return (EACCES);
/* Precedence goes to privilege. */
if (error1 == EPERM || error2 == EPERM)
return (EPERM);
/* Precedence goes to error over success; otherwise, arbitrary. */
if (error1 != 0)
return (error1);
return (error2);
}
Abstract access to the mbuf header label behind a new function, mbuf_to_label(). This permits the vast majority of entry point code to be unaware that labels are stored in m->m_pkthdr.label, such that we can experiment storage of labels elsewhere (such as in m_tags). Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 18:11:18 +00:00
static struct label *
mbuf_to_label(struct mbuf *mbuf)
{
Move MAC label storage for mbufs into m_tags from the m_pkthdr structure, returning some additional room in the first mbuf in a chain, and avoiding feature-specific contents in the mbuf header. To do this: - Modify mbuf_to_label() to extract the tag, returning NULL if not found. - Introduce mac_init_mbuf_tag() which does most of the work mac_init_mbuf() used to do, except on an m_tag rather than an mbuf. - Scale back mac_init_mbuf() to perform m_tag allocation and invoke mac_init_mbuf_tag(). - Replace mac_destroy_mbuf() with mac_destroy_mbuf_tag(), since m_tag's are now GC'd deep in the m_tag/mbuf code rather than at a higher level when mbufs are directly free()'d. - Add mac_copy_mbuf_tag() to support m_copy_pkthdr() and related notions. - Generally change all references to mbuf labels so that they use mbuf_to_label() rather than &mbuf->m_pkthdr.label. This required no changes in the MAC policies (yay!). - Tweak mbuf release routines to not call mac_destroy_mbuf(), tag destruction takes care of it for us now. - Remove MAC magic from m_copy_pkthdr() and m_move_pkthdr() -- the existing m_tag support does all this for us. Note that we can no longer just zero the m_tag list on the target mbuf, rather, we have to delete the chain because m_tag's will already be hung off freshly allocated mbuf's. - Tweak m_tag copying routines so that if we're copying a MAC m_tag, we don't do a binary copy, rather, we initialize the new storage and do a deep copy of the label. - Remove use of MAC_FLAG_INITIALIZED in a few bizarre places having to do with mbuf header copies previously. - When an mbuf is copied in ip_input(), we no longer need to explicitly copy the label because it will get handled by the m_tag code now. - No longer any weird handling of MAC labels in if_loop.c during header copies. - Add MPC_LOADTIME_FLAG_LABELMBUFS flag to Biba, MLS, mac_test. In mac_test, handle the label==NULL case, since it can be dynamically loaded. In order to improve performance with this change, introduce the notion of "lazy MAC label allocation" -- only allocate m_tag storage for MAC labels if we're running with a policy that uses MAC labels on mbufs. Policies declare this intent by setting the MPC_LOADTIME_FLAG_LABELMBUFS flag in their load-time flags field during declaration. Note: this opens up the possibility of post-boot policy modules getting back NULL slot entries even though they have policy invariants of non-NULL slot entries, as the policy might have been loaded after the mbuf was allocated, leaving the mbuf without label storage. Policies that cannot handle this case must be declared as NOTLATE, or must be modified. - mac_labelmbufs holds the current cumulative status as to whether any policies require mbuf labeling or not. This is updated whenever the active policy set changes by the function mac_policy_updateflags(). The function iterates the list and checks whether any have the flag set. Write access to this variable is protected by the policy list; read access is currently not protected for performance reasons. This might change if it causes problems. - Add MAC_POLICY_LIST_ASSERT_EXCLUSIVE() to permit the flags update function to assert appropriate locks. - This makes allocation in mac_init_mbuf() conditional on the flag. Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 20:39:06 +00:00
struct m_tag *tag;
Abstract access to the mbuf header label behind a new function, mbuf_to_label(). This permits the vast majority of entry point code to be unaware that labels are stored in m->m_pkthdr.label, such that we can experiment storage of labels elsewhere (such as in m_tags). Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 18:11:18 +00:00
struct label *label;
Move MAC label storage for mbufs into m_tags from the m_pkthdr structure, returning some additional room in the first mbuf in a chain, and avoiding feature-specific contents in the mbuf header. To do this: - Modify mbuf_to_label() to extract the tag, returning NULL if not found. - Introduce mac_init_mbuf_tag() which does most of the work mac_init_mbuf() used to do, except on an m_tag rather than an mbuf. - Scale back mac_init_mbuf() to perform m_tag allocation and invoke mac_init_mbuf_tag(). - Replace mac_destroy_mbuf() with mac_destroy_mbuf_tag(), since m_tag's are now GC'd deep in the m_tag/mbuf code rather than at a higher level when mbufs are directly free()'d. - Add mac_copy_mbuf_tag() to support m_copy_pkthdr() and related notions. - Generally change all references to mbuf labels so that they use mbuf_to_label() rather than &mbuf->m_pkthdr.label. This required no changes in the MAC policies (yay!). - Tweak mbuf release routines to not call mac_destroy_mbuf(), tag destruction takes care of it for us now. - Remove MAC magic from m_copy_pkthdr() and m_move_pkthdr() -- the existing m_tag support does all this for us. Note that we can no longer just zero the m_tag list on the target mbuf, rather, we have to delete the chain because m_tag's will already be hung off freshly allocated mbuf's. - Tweak m_tag copying routines so that if we're copying a MAC m_tag, we don't do a binary copy, rather, we initialize the new storage and do a deep copy of the label. - Remove use of MAC_FLAG_INITIALIZED in a few bizarre places having to do with mbuf header copies previously. - When an mbuf is copied in ip_input(), we no longer need to explicitly copy the label because it will get handled by the m_tag code now. - No longer any weird handling of MAC labels in if_loop.c during header copies. - Add MPC_LOADTIME_FLAG_LABELMBUFS flag to Biba, MLS, mac_test. In mac_test, handle the label==NULL case, since it can be dynamically loaded. In order to improve performance with this change, introduce the notion of "lazy MAC label allocation" -- only allocate m_tag storage for MAC labels if we're running with a policy that uses MAC labels on mbufs. Policies declare this intent by setting the MPC_LOADTIME_FLAG_LABELMBUFS flag in their load-time flags field during declaration. Note: this opens up the possibility of post-boot policy modules getting back NULL slot entries even though they have policy invariants of non-NULL slot entries, as the policy might have been loaded after the mbuf was allocated, leaving the mbuf without label storage. Policies that cannot handle this case must be declared as NOTLATE, or must be modified. - mac_labelmbufs holds the current cumulative status as to whether any policies require mbuf labeling or not. This is updated whenever the active policy set changes by the function mac_policy_updateflags(). The function iterates the list and checks whether any have the flag set. Write access to this variable is protected by the policy list; read access is currently not protected for performance reasons. This might change if it causes problems. - Add MAC_POLICY_LIST_ASSERT_EXCLUSIVE() to permit the flags update function to assert appropriate locks. - This makes allocation in mac_init_mbuf() conditional on the flag. Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 20:39:06 +00:00
tag = m_tag_find(mbuf, PACKET_TAG_MACLABEL, NULL);
label = (struct label *)(tag+1);
Abstract access to the mbuf header label behind a new function, mbuf_to_label(). This permits the vast majority of entry point code to be unaware that labels are stored in m->m_pkthdr.label, such that we can experiment storage of labels elsewhere (such as in m_tags). Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 18:11:18 +00:00
return (label);
}
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
static void
mac_init_label(struct label *label)
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
{
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
bzero(label, sizeof(*label));
label->l_flags = MAC_FLAG_INITIALIZED;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
static void
mac_destroy_label(struct label *label)
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
{
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
KASSERT(label->l_flags & MAC_FLAG_INITIALIZED,
("destroying uninitialized label"));
bzero(label, sizeof(*label));
/* implicit: label->l_flags &= ~MAC_FLAG_INITIALIZED; */
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
void
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
mac_init_bpfdesc(struct bpf_d *bpf_d)
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
{
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
mac_init_label(&bpf_d->bd_label);
MAC_PERFORM(init_bpfdesc_label, &bpf_d->bd_label);
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
#ifdef MAC_DEBUG
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
atomic_add_int(&nmacbpfdescs, 1);
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
#endif
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
static void
mac_init_cred_label(struct label *label)
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
{
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
mac_init_label(label);
MAC_PERFORM(init_cred_label, label);
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
#ifdef MAC_DEBUG
atomic_add_int(&nmaccreds, 1);
#endif
}
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
void
mac_init_cred(struct ucred *cred)
{
mac_init_cred_label(&cred->cr_label);
}
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
void
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
mac_init_devfsdirent(struct devfs_dirent *de)
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
{
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
mac_init_label(&de->de_label);
MAC_PERFORM(init_devfsdirent_label, &de->de_label);
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
#ifdef MAC_DEBUG
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
atomic_add_int(&nmacdevfsdirents, 1);
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
#endif
}
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
static void
mac_init_ifnet_label(struct label *label)
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
{
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
mac_init_label(label);
MAC_PERFORM(init_ifnet_label, label);
Push 'security.mac.debug_label_fallback' behind options MAC_DEBUG. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-10-01 03:24:20 +00:00
#ifdef MAC_DEBUG
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
atomic_add_int(&nmacifnets, 1);
Push 'security.mac.debug_label_fallback' behind options MAC_DEBUG. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-10-01 03:24:20 +00:00
#endif
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
}
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
void
mac_init_ifnet(struct ifnet *ifp)
{
mac_init_ifnet_label(&ifp->if_label);
}
Modify the mac_init_ipq() MAC Framework entry point to accept an additional flags argument to indicate blocking disposition, and pass in M_NOWAIT from the IP reassembly code to indicate that blocking is not OK when labeling a new IP fragment reassembly queue. This should eliminate some of the WITNESS warnings that have started popping up since fine-grained IP stack locking started going in; if memory allocation fails, the creation of the fragment queue will be aborted. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-03-26 15:12:03 +00:00
int
mac_init_ipq(struct ipq *ipq, int flag)
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
{
Modify the mac_init_ipq() MAC Framework entry point to accept an additional flags argument to indicate blocking disposition, and pass in M_NOWAIT from the IP reassembly code to indicate that blocking is not OK when labeling a new IP fragment reassembly queue. This should eliminate some of the WITNESS warnings that have started popping up since fine-grained IP stack locking started going in; if memory allocation fails, the creation of the fragment queue will be aborted. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-03-26 15:12:03 +00:00
int error;
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
mac_init_label(&ipq->ipq_label);
Modify the mac_init_ipq() MAC Framework entry point to accept an additional flags argument to indicate blocking disposition, and pass in M_NOWAIT from the IP reassembly code to indicate that blocking is not OK when labeling a new IP fragment reassembly queue. This should eliminate some of the WITNESS warnings that have started popping up since fine-grained IP stack locking started going in; if memory allocation fails, the creation of the fragment queue will be aborted. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-03-26 15:12:03 +00:00
MAC_CHECK(init_ipq_label, &ipq->ipq_label, flag);
if (error) {
MAC_PERFORM(destroy_ipq_label, &ipq->ipq_label);
mac_destroy_label(&ipq->ipq_label);
}
Push 'security.mac.debug_label_fallback' behind options MAC_DEBUG. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-10-01 03:24:20 +00:00
#ifdef MAC_DEBUG
Modify the mac_init_ipq() MAC Framework entry point to accept an additional flags argument to indicate blocking disposition, and pass in M_NOWAIT from the IP reassembly code to indicate that blocking is not OK when labeling a new IP fragment reassembly queue. This should eliminate some of the WITNESS warnings that have started popping up since fine-grained IP stack locking started going in; if memory allocation fails, the creation of the fragment queue will be aborted. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-03-26 15:12:03 +00:00
if (error == 0)
atomic_add_int(&nmacipqs, 1);
Push 'security.mac.debug_label_fallback' behind options MAC_DEBUG. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-10-01 03:24:20 +00:00
#endif
Modify the mac_init_ipq() MAC Framework entry point to accept an additional flags argument to indicate blocking disposition, and pass in M_NOWAIT from the IP reassembly code to indicate that blocking is not OK when labeling a new IP fragment reassembly queue. This should eliminate some of the WITNESS warnings that have started popping up since fine-grained IP stack locking started going in; if memory allocation fails, the creation of the fragment queue will be aborted. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-03-26 15:12:03 +00:00
return (error);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
int
Move MAC label storage for mbufs into m_tags from the m_pkthdr structure, returning some additional room in the first mbuf in a chain, and avoiding feature-specific contents in the mbuf header. To do this: - Modify mbuf_to_label() to extract the tag, returning NULL if not found. - Introduce mac_init_mbuf_tag() which does most of the work mac_init_mbuf() used to do, except on an m_tag rather than an mbuf. - Scale back mac_init_mbuf() to perform m_tag allocation and invoke mac_init_mbuf_tag(). - Replace mac_destroy_mbuf() with mac_destroy_mbuf_tag(), since m_tag's are now GC'd deep in the m_tag/mbuf code rather than at a higher level when mbufs are directly free()'d. - Add mac_copy_mbuf_tag() to support m_copy_pkthdr() and related notions. - Generally change all references to mbuf labels so that they use mbuf_to_label() rather than &mbuf->m_pkthdr.label. This required no changes in the MAC policies (yay!). - Tweak mbuf release routines to not call mac_destroy_mbuf(), tag destruction takes care of it for us now. - Remove MAC magic from m_copy_pkthdr() and m_move_pkthdr() -- the existing m_tag support does all this for us. Note that we can no longer just zero the m_tag list on the target mbuf, rather, we have to delete the chain because m_tag's will already be hung off freshly allocated mbuf's. - Tweak m_tag copying routines so that if we're copying a MAC m_tag, we don't do a binary copy, rather, we initialize the new storage and do a deep copy of the label. - Remove use of MAC_FLAG_INITIALIZED in a few bizarre places having to do with mbuf header copies previously. - When an mbuf is copied in ip_input(), we no longer need to explicitly copy the label because it will get handled by the m_tag code now. - No longer any weird handling of MAC labels in if_loop.c during header copies. - Add MPC_LOADTIME_FLAG_LABELMBUFS flag to Biba, MLS, mac_test. In mac_test, handle the label==NULL case, since it can be dynamically loaded. In order to improve performance with this change, introduce the notion of "lazy MAC label allocation" -- only allocate m_tag storage for MAC labels if we're running with a policy that uses MAC labels on mbufs. Policies declare this intent by setting the MPC_LOADTIME_FLAG_LABELMBUFS flag in their load-time flags field during declaration. Note: this opens up the possibility of post-boot policy modules getting back NULL slot entries even though they have policy invariants of non-NULL slot entries, as the policy might have been loaded after the mbuf was allocated, leaving the mbuf without label storage. Policies that cannot handle this case must be declared as NOTLATE, or must be modified. - mac_labelmbufs holds the current cumulative status as to whether any policies require mbuf labeling or not. This is updated whenever the active policy set changes by the function mac_policy_updateflags(). The function iterates the list and checks whether any have the flag set. Write access to this variable is protected by the policy list; read access is currently not protected for performance reasons. This might change if it causes problems. - Add MAC_POLICY_LIST_ASSERT_EXCLUSIVE() to permit the flags update function to assert appropriate locks. - This makes allocation in mac_init_mbuf() conditional on the flag. Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 20:39:06 +00:00
mac_init_mbuf_tag(struct m_tag *tag, int flag)
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
{
Move MAC label storage for mbufs into m_tags from the m_pkthdr structure, returning some additional room in the first mbuf in a chain, and avoiding feature-specific contents in the mbuf header. To do this: - Modify mbuf_to_label() to extract the tag, returning NULL if not found. - Introduce mac_init_mbuf_tag() which does most of the work mac_init_mbuf() used to do, except on an m_tag rather than an mbuf. - Scale back mac_init_mbuf() to perform m_tag allocation and invoke mac_init_mbuf_tag(). - Replace mac_destroy_mbuf() with mac_destroy_mbuf_tag(), since m_tag's are now GC'd deep in the m_tag/mbuf code rather than at a higher level when mbufs are directly free()'d. - Add mac_copy_mbuf_tag() to support m_copy_pkthdr() and related notions. - Generally change all references to mbuf labels so that they use mbuf_to_label() rather than &mbuf->m_pkthdr.label. This required no changes in the MAC policies (yay!). - Tweak mbuf release routines to not call mac_destroy_mbuf(), tag destruction takes care of it for us now. - Remove MAC magic from m_copy_pkthdr() and m_move_pkthdr() -- the existing m_tag support does all this for us. Note that we can no longer just zero the m_tag list on the target mbuf, rather, we have to delete the chain because m_tag's will already be hung off freshly allocated mbuf's. - Tweak m_tag copying routines so that if we're copying a MAC m_tag, we don't do a binary copy, rather, we initialize the new storage and do a deep copy of the label. - Remove use of MAC_FLAG_INITIALIZED in a few bizarre places having to do with mbuf header copies previously. - When an mbuf is copied in ip_input(), we no longer need to explicitly copy the label because it will get handled by the m_tag code now. - No longer any weird handling of MAC labels in if_loop.c during header copies. - Add MPC_LOADTIME_FLAG_LABELMBUFS flag to Biba, MLS, mac_test. In mac_test, handle the label==NULL case, since it can be dynamically loaded. In order to improve performance with this change, introduce the notion of "lazy MAC label allocation" -- only allocate m_tag storage for MAC labels if we're running with a policy that uses MAC labels on mbufs. Policies declare this intent by setting the MPC_LOADTIME_FLAG_LABELMBUFS flag in their load-time flags field during declaration. Note: this opens up the possibility of post-boot policy modules getting back NULL slot entries even though they have policy invariants of non-NULL slot entries, as the policy might have been loaded after the mbuf was allocated, leaving the mbuf without label storage. Policies that cannot handle this case must be declared as NOTLATE, or must be modified. - mac_labelmbufs holds the current cumulative status as to whether any policies require mbuf labeling or not. This is updated whenever the active policy set changes by the function mac_policy_updateflags(). The function iterates the list and checks whether any have the flag set. Write access to this variable is protected by the policy list; read access is currently not protected for performance reasons. This might change if it causes problems. - Add MAC_POLICY_LIST_ASSERT_EXCLUSIVE() to permit the flags update function to assert appropriate locks. - This makes allocation in mac_init_mbuf() conditional on the flag. Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 20:39:06 +00:00
struct label *label;
mac_init_mbuf_tag() accepts malloc flags, not mbuf allocator flags, so don't try and convert the argument flags to malloc flags, or we risk implicitly requesting blocking and generating witness warnings. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-15 19:33:23 +00:00
int error;
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
Move MAC label storage for mbufs into m_tags from the m_pkthdr structure, returning some additional room in the first mbuf in a chain, and avoiding feature-specific contents in the mbuf header. To do this: - Modify mbuf_to_label() to extract the tag, returning NULL if not found. - Introduce mac_init_mbuf_tag() which does most of the work mac_init_mbuf() used to do, except on an m_tag rather than an mbuf. - Scale back mac_init_mbuf() to perform m_tag allocation and invoke mac_init_mbuf_tag(). - Replace mac_destroy_mbuf() with mac_destroy_mbuf_tag(), since m_tag's are now GC'd deep in the m_tag/mbuf code rather than at a higher level when mbufs are directly free()'d. - Add mac_copy_mbuf_tag() to support m_copy_pkthdr() and related notions. - Generally change all references to mbuf labels so that they use mbuf_to_label() rather than &mbuf->m_pkthdr.label. This required no changes in the MAC policies (yay!). - Tweak mbuf release routines to not call mac_destroy_mbuf(), tag destruction takes care of it for us now. - Remove MAC magic from m_copy_pkthdr() and m_move_pkthdr() -- the existing m_tag support does all this for us. Note that we can no longer just zero the m_tag list on the target mbuf, rather, we have to delete the chain because m_tag's will already be hung off freshly allocated mbuf's. - Tweak m_tag copying routines so that if we're copying a MAC m_tag, we don't do a binary copy, rather, we initialize the new storage and do a deep copy of the label. - Remove use of MAC_FLAG_INITIALIZED in a few bizarre places having to do with mbuf header copies previously. - When an mbuf is copied in ip_input(), we no longer need to explicitly copy the label because it will get handled by the m_tag code now. - No longer any weird handling of MAC labels in if_loop.c during header copies. - Add MPC_LOADTIME_FLAG_LABELMBUFS flag to Biba, MLS, mac_test. In mac_test, handle the label==NULL case, since it can be dynamically loaded. In order to improve performance with this change, introduce the notion of "lazy MAC label allocation" -- only allocate m_tag storage for MAC labels if we're running with a policy that uses MAC labels on mbufs. Policies declare this intent by setting the MPC_LOADTIME_FLAG_LABELMBUFS flag in their load-time flags field during declaration. Note: this opens up the possibility of post-boot policy modules getting back NULL slot entries even though they have policy invariants of non-NULL slot entries, as the policy might have been loaded after the mbuf was allocated, leaving the mbuf without label storage. Policies that cannot handle this case must be declared as NOTLATE, or must be modified. - mac_labelmbufs holds the current cumulative status as to whether any policies require mbuf labeling or not. This is updated whenever the active policy set changes by the function mac_policy_updateflags(). The function iterates the list and checks whether any have the flag set. Write access to this variable is protected by the policy list; read access is currently not protected for performance reasons. This might change if it causes problems. - Add MAC_POLICY_LIST_ASSERT_EXCLUSIVE() to permit the flags update function to assert appropriate locks. - This makes allocation in mac_init_mbuf() conditional on the flag. Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 20:39:06 +00:00
label = (struct label *) (tag + 1);
mac_init_label(label);
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
mac_init_mbuf_tag() accepts malloc flags, not mbuf allocator flags, so don't try and convert the argument flags to malloc flags, or we risk implicitly requesting blocking and generating witness warnings. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-15 19:33:23 +00:00
MAC_CHECK(init_mbuf_label, label, flag);
While the MAC API has supported the ability to handle M_NOWAIT passed to mbuf label initialization, that functionality was never merged to the main tree. Go ahead and merge that functionality now. Note that this requires policy modules to accept the case where the label element may be destroyed even if init has not succeeded on it (in the event that policy failed the init). This will shortly also apply to sockets. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:44:49 +00:00
if (error) {
Move MAC label storage for mbufs into m_tags from the m_pkthdr structure, returning some additional room in the first mbuf in a chain, and avoiding feature-specific contents in the mbuf header. To do this: - Modify mbuf_to_label() to extract the tag, returning NULL if not found. - Introduce mac_init_mbuf_tag() which does most of the work mac_init_mbuf() used to do, except on an m_tag rather than an mbuf. - Scale back mac_init_mbuf() to perform m_tag allocation and invoke mac_init_mbuf_tag(). - Replace mac_destroy_mbuf() with mac_destroy_mbuf_tag(), since m_tag's are now GC'd deep in the m_tag/mbuf code rather than at a higher level when mbufs are directly free()'d. - Add mac_copy_mbuf_tag() to support m_copy_pkthdr() and related notions. - Generally change all references to mbuf labels so that they use mbuf_to_label() rather than &mbuf->m_pkthdr.label. This required no changes in the MAC policies (yay!). - Tweak mbuf release routines to not call mac_destroy_mbuf(), tag destruction takes care of it for us now. - Remove MAC magic from m_copy_pkthdr() and m_move_pkthdr() -- the existing m_tag support does all this for us. Note that we can no longer just zero the m_tag list on the target mbuf, rather, we have to delete the chain because m_tag's will already be hung off freshly allocated mbuf's. - Tweak m_tag copying routines so that if we're copying a MAC m_tag, we don't do a binary copy, rather, we initialize the new storage and do a deep copy of the label. - Remove use of MAC_FLAG_INITIALIZED in a few bizarre places having to do with mbuf header copies previously. - When an mbuf is copied in ip_input(), we no longer need to explicitly copy the label because it will get handled by the m_tag code now. - No longer any weird handling of MAC labels in if_loop.c during header copies. - Add MPC_LOADTIME_FLAG_LABELMBUFS flag to Biba, MLS, mac_test. In mac_test, handle the label==NULL case, since it can be dynamically loaded. In order to improve performance with this change, introduce the notion of "lazy MAC label allocation" -- only allocate m_tag storage for MAC labels if we're running with a policy that uses MAC labels on mbufs. Policies declare this intent by setting the MPC_LOADTIME_FLAG_LABELMBUFS flag in their load-time flags field during declaration. Note: this opens up the possibility of post-boot policy modules getting back NULL slot entries even though they have policy invariants of non-NULL slot entries, as the policy might have been loaded after the mbuf was allocated, leaving the mbuf without label storage. Policies that cannot handle this case must be declared as NOTLATE, or must be modified. - mac_labelmbufs holds the current cumulative status as to whether any policies require mbuf labeling or not. This is updated whenever the active policy set changes by the function mac_policy_updateflags(). The function iterates the list and checks whether any have the flag set. Write access to this variable is protected by the policy list; read access is currently not protected for performance reasons. This might change if it causes problems. - Add MAC_POLICY_LIST_ASSERT_EXCLUSIVE() to permit the flags update function to assert appropriate locks. - This makes allocation in mac_init_mbuf() conditional on the flag. Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 20:39:06 +00:00
MAC_PERFORM(destroy_mbuf_label, label);
mac_destroy_label(label);
While the MAC API has supported the ability to handle M_NOWAIT passed to mbuf label initialization, that functionality was never merged to the main tree. Go ahead and merge that functionality now. Note that this requires policy modules to accept the case where the label element may be destroyed even if init has not succeeded on it (in the event that policy failed the init). This will shortly also apply to sockets. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:44:49 +00:00
}
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
#ifdef MAC_DEBUG
While the MAC API has supported the ability to handle M_NOWAIT passed to mbuf label initialization, that functionality was never merged to the main tree. Go ahead and merge that functionality now. Note that this requires policy modules to accept the case where the label element may be destroyed even if init has not succeeded on it (in the event that policy failed the init). This will shortly also apply to sockets. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:44:49 +00:00
if (error == 0)
atomic_add_int(&nmacmbufs, 1);
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
#endif
While the MAC API has supported the ability to handle M_NOWAIT passed to mbuf label initialization, that functionality was never merged to the main tree. Go ahead and merge that functionality now. Note that this requires policy modules to accept the case where the label element may be destroyed even if init has not succeeded on it (in the event that policy failed the init). This will shortly also apply to sockets. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:44:49 +00:00
return (error);
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
}
Move MAC label storage for mbufs into m_tags from the m_pkthdr structure, returning some additional room in the first mbuf in a chain, and avoiding feature-specific contents in the mbuf header. To do this: - Modify mbuf_to_label() to extract the tag, returning NULL if not found. - Introduce mac_init_mbuf_tag() which does most of the work mac_init_mbuf() used to do, except on an m_tag rather than an mbuf. - Scale back mac_init_mbuf() to perform m_tag allocation and invoke mac_init_mbuf_tag(). - Replace mac_destroy_mbuf() with mac_destroy_mbuf_tag(), since m_tag's are now GC'd deep in the m_tag/mbuf code rather than at a higher level when mbufs are directly free()'d. - Add mac_copy_mbuf_tag() to support m_copy_pkthdr() and related notions. - Generally change all references to mbuf labels so that they use mbuf_to_label() rather than &mbuf->m_pkthdr.label. This required no changes in the MAC policies (yay!). - Tweak mbuf release routines to not call mac_destroy_mbuf(), tag destruction takes care of it for us now. - Remove MAC magic from m_copy_pkthdr() and m_move_pkthdr() -- the existing m_tag support does all this for us. Note that we can no longer just zero the m_tag list on the target mbuf, rather, we have to delete the chain because m_tag's will already be hung off freshly allocated mbuf's. - Tweak m_tag copying routines so that if we're copying a MAC m_tag, we don't do a binary copy, rather, we initialize the new storage and do a deep copy of the label. - Remove use of MAC_FLAG_INITIALIZED in a few bizarre places having to do with mbuf header copies previously. - When an mbuf is copied in ip_input(), we no longer need to explicitly copy the label because it will get handled by the m_tag code now. - No longer any weird handling of MAC labels in if_loop.c during header copies. - Add MPC_LOADTIME_FLAG_LABELMBUFS flag to Biba, MLS, mac_test. In mac_test, handle the label==NULL case, since it can be dynamically loaded. In order to improve performance with this change, introduce the notion of "lazy MAC label allocation" -- only allocate m_tag storage for MAC labels if we're running with a policy that uses MAC labels on mbufs. Policies declare this intent by setting the MPC_LOADTIME_FLAG_LABELMBUFS flag in their load-time flags field during declaration. Note: this opens up the possibility of post-boot policy modules getting back NULL slot entries even though they have policy invariants of non-NULL slot entries, as the policy might have been loaded after the mbuf was allocated, leaving the mbuf without label storage. Policies that cannot handle this case must be declared as NOTLATE, or must be modified. - mac_labelmbufs holds the current cumulative status as to whether any policies require mbuf labeling or not. This is updated whenever the active policy set changes by the function mac_policy_updateflags(). The function iterates the list and checks whether any have the flag set. Write access to this variable is protected by the policy list; read access is currently not protected for performance reasons. This might change if it causes problems. - Add MAC_POLICY_LIST_ASSERT_EXCLUSIVE() to permit the flags update function to assert appropriate locks. - This makes allocation in mac_init_mbuf() conditional on the flag. Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 20:39:06 +00:00
int
mac_init_mbuf(struct mbuf *m, int flag)
{
struct m_tag *tag;
int error;
M_ASSERTPKTHDR(m);
#ifndef MAC_ALWAYS_LABEL_MBUF
/*
Attempt to simplify #ifdef logic for MAC_ALWAYS_LABEL_MBUF. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-08-01 15:45:14 +00:00
* If conditionally allocating mbuf labels, don't allocate unless
* they are required.
Move MAC label storage for mbufs into m_tags from the m_pkthdr structure, returning some additional room in the first mbuf in a chain, and avoiding feature-specific contents in the mbuf header. To do this: - Modify mbuf_to_label() to extract the tag, returning NULL if not found. - Introduce mac_init_mbuf_tag() which does most of the work mac_init_mbuf() used to do, except on an m_tag rather than an mbuf. - Scale back mac_init_mbuf() to perform m_tag allocation and invoke mac_init_mbuf_tag(). - Replace mac_destroy_mbuf() with mac_destroy_mbuf_tag(), since m_tag's are now GC'd deep in the m_tag/mbuf code rather than at a higher level when mbufs are directly free()'d. - Add mac_copy_mbuf_tag() to support m_copy_pkthdr() and related notions. - Generally change all references to mbuf labels so that they use mbuf_to_label() rather than &mbuf->m_pkthdr.label. This required no changes in the MAC policies (yay!). - Tweak mbuf release routines to not call mac_destroy_mbuf(), tag destruction takes care of it for us now. - Remove MAC magic from m_copy_pkthdr() and m_move_pkthdr() -- the existing m_tag support does all this for us. Note that we can no longer just zero the m_tag list on the target mbuf, rather, we have to delete the chain because m_tag's will already be hung off freshly allocated mbuf's. - Tweak m_tag copying routines so that if we're copying a MAC m_tag, we don't do a binary copy, rather, we initialize the new storage and do a deep copy of the label. - Remove use of MAC_FLAG_INITIALIZED in a few bizarre places having to do with mbuf header copies previously. - When an mbuf is copied in ip_input(), we no longer need to explicitly copy the label because it will get handled by the m_tag code now. - No longer any weird handling of MAC labels in if_loop.c during header copies. - Add MPC_LOADTIME_FLAG_LABELMBUFS flag to Biba, MLS, mac_test. In mac_test, handle the label==NULL case, since it can be dynamically loaded. In order to improve performance with this change, introduce the notion of "lazy MAC label allocation" -- only allocate m_tag storage for MAC labels if we're running with a policy that uses MAC labels on mbufs. Policies declare this intent by setting the MPC_LOADTIME_FLAG_LABELMBUFS flag in their load-time flags field during declaration. Note: this opens up the possibility of post-boot policy modules getting back NULL slot entries even though they have policy invariants of non-NULL slot entries, as the policy might have been loaded after the mbuf was allocated, leaving the mbuf without label storage. Policies that cannot handle this case must be declared as NOTLATE, or must be modified. - mac_labelmbufs holds the current cumulative status as to whether any policies require mbuf labeling or not. This is updated whenever the active policy set changes by the function mac_policy_updateflags(). The function iterates the list and checks whether any have the flag set. Write access to this variable is protected by the policy list; read access is currently not protected for performance reasons. This might change if it causes problems. - Add MAC_POLICY_LIST_ASSERT_EXCLUSIVE() to permit the flags update function to assert appropriate locks. - This makes allocation in mac_init_mbuf() conditional on the flag. Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 20:39:06 +00:00
*/
Attempt to simplify #ifdef logic for MAC_ALWAYS_LABEL_MBUF. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-08-01 15:45:14 +00:00
if (!mac_labelmbufs)
return (0);
Move MAC label storage for mbufs into m_tags from the m_pkthdr structure, returning some additional room in the first mbuf in a chain, and avoiding feature-specific contents in the mbuf header. To do this: - Modify mbuf_to_label() to extract the tag, returning NULL if not found. - Introduce mac_init_mbuf_tag() which does most of the work mac_init_mbuf() used to do, except on an m_tag rather than an mbuf. - Scale back mac_init_mbuf() to perform m_tag allocation and invoke mac_init_mbuf_tag(). - Replace mac_destroy_mbuf() with mac_destroy_mbuf_tag(), since m_tag's are now GC'd deep in the m_tag/mbuf code rather than at a higher level when mbufs are directly free()'d. - Add mac_copy_mbuf_tag() to support m_copy_pkthdr() and related notions. - Generally change all references to mbuf labels so that they use mbuf_to_label() rather than &mbuf->m_pkthdr.label. This required no changes in the MAC policies (yay!). - Tweak mbuf release routines to not call mac_destroy_mbuf(), tag destruction takes care of it for us now. - Remove MAC magic from m_copy_pkthdr() and m_move_pkthdr() -- the existing m_tag support does all this for us. Note that we can no longer just zero the m_tag list on the target mbuf, rather, we have to delete the chain because m_tag's will already be hung off freshly allocated mbuf's. - Tweak m_tag copying routines so that if we're copying a MAC m_tag, we don't do a binary copy, rather, we initialize the new storage and do a deep copy of the label. - Remove use of MAC_FLAG_INITIALIZED in a few bizarre places having to do with mbuf header copies previously. - When an mbuf is copied in ip_input(), we no longer need to explicitly copy the label because it will get handled by the m_tag code now. - No longer any weird handling of MAC labels in if_loop.c during header copies. - Add MPC_LOADTIME_FLAG_LABELMBUFS flag to Biba, MLS, mac_test. In mac_test, handle the label==NULL case, since it can be dynamically loaded. In order to improve performance with this change, introduce the notion of "lazy MAC label allocation" -- only allocate m_tag storage for MAC labels if we're running with a policy that uses MAC labels on mbufs. Policies declare this intent by setting the MPC_LOADTIME_FLAG_LABELMBUFS flag in their load-time flags field during declaration. Note: this opens up the possibility of post-boot policy modules getting back NULL slot entries even though they have policy invariants of non-NULL slot entries, as the policy might have been loaded after the mbuf was allocated, leaving the mbuf without label storage. Policies that cannot handle this case must be declared as NOTLATE, or must be modified. - mac_labelmbufs holds the current cumulative status as to whether any policies require mbuf labeling or not. This is updated whenever the active policy set changes by the function mac_policy_updateflags(). The function iterates the list and checks whether any have the flag set. Write access to this variable is protected by the policy list; read access is currently not protected for performance reasons. This might change if it causes problems. - Add MAC_POLICY_LIST_ASSERT_EXCLUSIVE() to permit the flags update function to assert appropriate locks. - This makes allocation in mac_init_mbuf() conditional on the flag. Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 20:39:06 +00:00
#endif
Attempt to simplify #ifdef logic for MAC_ALWAYS_LABEL_MBUF. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-08-01 15:45:14 +00:00
tag = m_tag_get(PACKET_TAG_MACLABEL, sizeof(struct label),
flag);
if (tag == NULL)
return (ENOMEM);
error = mac_init_mbuf_tag(tag, flag);
if (error) {
m_tag_free(tag);
return (error);
Move MAC label storage for mbufs into m_tags from the m_pkthdr structure, returning some additional room in the first mbuf in a chain, and avoiding feature-specific contents in the mbuf header. To do this: - Modify mbuf_to_label() to extract the tag, returning NULL if not found. - Introduce mac_init_mbuf_tag() which does most of the work mac_init_mbuf() used to do, except on an m_tag rather than an mbuf. - Scale back mac_init_mbuf() to perform m_tag allocation and invoke mac_init_mbuf_tag(). - Replace mac_destroy_mbuf() with mac_destroy_mbuf_tag(), since m_tag's are now GC'd deep in the m_tag/mbuf code rather than at a higher level when mbufs are directly free()'d. - Add mac_copy_mbuf_tag() to support m_copy_pkthdr() and related notions. - Generally change all references to mbuf labels so that they use mbuf_to_label() rather than &mbuf->m_pkthdr.label. This required no changes in the MAC policies (yay!). - Tweak mbuf release routines to not call mac_destroy_mbuf(), tag destruction takes care of it for us now. - Remove MAC magic from m_copy_pkthdr() and m_move_pkthdr() -- the existing m_tag support does all this for us. Note that we can no longer just zero the m_tag list on the target mbuf, rather, we have to delete the chain because m_tag's will already be hung off freshly allocated mbuf's. - Tweak m_tag copying routines so that if we're copying a MAC m_tag, we don't do a binary copy, rather, we initialize the new storage and do a deep copy of the label. - Remove use of MAC_FLAG_INITIALIZED in a few bizarre places having to do with mbuf header copies previously. - When an mbuf is copied in ip_input(), we no longer need to explicitly copy the label because it will get handled by the m_tag code now. - No longer any weird handling of MAC labels in if_loop.c during header copies. - Add MPC_LOADTIME_FLAG_LABELMBUFS flag to Biba, MLS, mac_test. In mac_test, handle the label==NULL case, since it can be dynamically loaded. In order to improve performance with this change, introduce the notion of "lazy MAC label allocation" -- only allocate m_tag storage for MAC labels if we're running with a policy that uses MAC labels on mbufs. Policies declare this intent by setting the MPC_LOADTIME_FLAG_LABELMBUFS flag in their load-time flags field during declaration. Note: this opens up the possibility of post-boot policy modules getting back NULL slot entries even though they have policy invariants of non-NULL slot entries, as the policy might have been loaded after the mbuf was allocated, leaving the mbuf without label storage. Policies that cannot handle this case must be declared as NOTLATE, or must be modified. - mac_labelmbufs holds the current cumulative status as to whether any policies require mbuf labeling or not. This is updated whenever the active policy set changes by the function mac_policy_updateflags(). The function iterates the list and checks whether any have the flag set. Write access to this variable is protected by the policy list; read access is currently not protected for performance reasons. This might change if it causes problems. - Add MAC_POLICY_LIST_ASSERT_EXCLUSIVE() to permit the flags update function to assert appropriate locks. - This makes allocation in mac_init_mbuf() conditional on the flag. Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 20:39:06 +00:00
}
Attempt to simplify #ifdef logic for MAC_ALWAYS_LABEL_MBUF. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-08-01 15:45:14 +00:00
m_tag_prepend(m, tag);
Move MAC label storage for mbufs into m_tags from the m_pkthdr structure, returning some additional room in the first mbuf in a chain, and avoiding feature-specific contents in the mbuf header. To do this: - Modify mbuf_to_label() to extract the tag, returning NULL if not found. - Introduce mac_init_mbuf_tag() which does most of the work mac_init_mbuf() used to do, except on an m_tag rather than an mbuf. - Scale back mac_init_mbuf() to perform m_tag allocation and invoke mac_init_mbuf_tag(). - Replace mac_destroy_mbuf() with mac_destroy_mbuf_tag(), since m_tag's are now GC'd deep in the m_tag/mbuf code rather than at a higher level when mbufs are directly free()'d. - Add mac_copy_mbuf_tag() to support m_copy_pkthdr() and related notions. - Generally change all references to mbuf labels so that they use mbuf_to_label() rather than &mbuf->m_pkthdr.label. This required no changes in the MAC policies (yay!). - Tweak mbuf release routines to not call mac_destroy_mbuf(), tag destruction takes care of it for us now. - Remove MAC magic from m_copy_pkthdr() and m_move_pkthdr() -- the existing m_tag support does all this for us. Note that we can no longer just zero the m_tag list on the target mbuf, rather, we have to delete the chain because m_tag's will already be hung off freshly allocated mbuf's. - Tweak m_tag copying routines so that if we're copying a MAC m_tag, we don't do a binary copy, rather, we initialize the new storage and do a deep copy of the label. - Remove use of MAC_FLAG_INITIALIZED in a few bizarre places having to do with mbuf header copies previously. - When an mbuf is copied in ip_input(), we no longer need to explicitly copy the label because it will get handled by the m_tag code now. - No longer any weird handling of MAC labels in if_loop.c during header copies. - Add MPC_LOADTIME_FLAG_LABELMBUFS flag to Biba, MLS, mac_test. In mac_test, handle the label==NULL case, since it can be dynamically loaded. In order to improve performance with this change, introduce the notion of "lazy MAC label allocation" -- only allocate m_tag storage for MAC labels if we're running with a policy that uses MAC labels on mbufs. Policies declare this intent by setting the MPC_LOADTIME_FLAG_LABELMBUFS flag in their load-time flags field during declaration. Note: this opens up the possibility of post-boot policy modules getting back NULL slot entries even though they have policy invariants of non-NULL slot entries, as the policy might have been loaded after the mbuf was allocated, leaving the mbuf without label storage. Policies that cannot handle this case must be declared as NOTLATE, or must be modified. - mac_labelmbufs holds the current cumulative status as to whether any policies require mbuf labeling or not. This is updated whenever the active policy set changes by the function mac_policy_updateflags(). The function iterates the list and checks whether any have the flag set. Write access to this variable is protected by the policy list; read access is currently not protected for performance reasons. This might change if it causes problems. - Add MAC_POLICY_LIST_ASSERT_EXCLUSIVE() to permit the flags update function to assert appropriate locks. - This makes allocation in mac_init_mbuf() conditional on the flag. Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 20:39:06 +00:00
return (0);
}
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
void
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
mac_init_mount(struct mount *mp)
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
{
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
mac_init_label(&mp->mnt_mntlabel);
mac_init_label(&mp->mnt_fslabel);
MAC_PERFORM(init_mount_label, &mp->mnt_mntlabel);
MAC_PERFORM(init_mount_fs_label, &mp->mnt_fslabel);
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
#ifdef MAC_DEBUG
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
atomic_add_int(&nmacmounts, 1);
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
#endif
}
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
static void
mac_init_pipe_label(struct label *label)
{
mac_init_label(label);
MAC_PERFORM(init_pipe_label, label);
#ifdef MAC_DEBUG
atomic_add_int(&nmacpipes, 1);
#endif
}
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
void
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
mac_init_pipe(struct pipe *pipe)
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
{
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
struct label *label;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
label = malloc(sizeof(struct label), M_MACPIPELABEL, M_ZERO|M_WAITOK);
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
pipe->pipe_label = label;
pipe->pipe_peer->pipe_label = label;
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
mac_init_pipe_label(label);
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
}
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
Introduce p_label, extensible security label storage for the MAC framework in struct proc. While the process label is actually stored in the struct ucred pointed to by p_ucred, there is a need for transient storage that may be used when asynchronous (deferred) updates need to be performed on the "real" label for locking reasons. Unlike other label storage, this label has no locking semantics, relying on policies to provide their own protection for the label contents, meaning that a policy leaf mutex may be used, avoiding lock order issues. This permits policies that act based on historical process behavior (such as audit policies, the MAC Framework port of LOMAC, etc) can update process properties even when many existing locks are held without violating the lock order. No currently committed policies implement use of this label storage. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-11-20 15:41:25 +00:00
void
mac_init_proc(struct proc *p)
{
mac_init_label(&p->p_label);
MAC_PERFORM(init_proc_label, &p->p_label);
#ifdef MAC_DEBUG
atomic_add_int(&nmacprocs, 1);
#endif
}
Modify label allocation semantics for sockets: pass in soalloc's malloc flags so that we can call malloc with M_NOWAIT if necessary, avoiding potential sleeps while holding mutexes in the TCP syncache code. Similar to the existing support for mbuf label allocation: if we can't allocate all the necessary label store in each policy, we back out the label allocation and fail the socket creation. Sync from MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 21:23:47 +00:00
static int
mac_init_socket_label(struct label *label, int flag)
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
{
Modify label allocation semantics for sockets: pass in soalloc's malloc flags so that we can call malloc with M_NOWAIT if necessary, avoiding potential sleeps while holding mutexes in the TCP syncache code. Similar to the existing support for mbuf label allocation: if we can't allocate all the necessary label store in each policy, we back out the label allocation and fail the socket creation. Sync from MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 21:23:47 +00:00
int error;
mac_init_label(label);
MAC_CHECK(init_socket_label, label, flag);
if (error) {
MAC_PERFORM(destroy_socket_label, label);
mac_destroy_label(label);
}
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
#ifdef MAC_DEBUG
Modify label allocation semantics for sockets: pass in soalloc's malloc flags so that we can call malloc with M_NOWAIT if necessary, avoiding potential sleeps while holding mutexes in the TCP syncache code. Similar to the existing support for mbuf label allocation: if we can't allocate all the necessary label store in each policy, we back out the label allocation and fail the socket creation. Sync from MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 21:23:47 +00:00
if (error == 0)
atomic_add_int(&nmacsockets, 1);
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
#endif
Modify label allocation semantics for sockets: pass in soalloc's malloc flags so that we can call malloc with M_NOWAIT if necessary, avoiding potential sleeps while holding mutexes in the TCP syncache code. Similar to the existing support for mbuf label allocation: if we can't allocate all the necessary label store in each policy, we back out the label allocation and fail the socket creation. Sync from MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 21:23:47 +00:00
return (error);
}
static int
mac_init_socket_peer_label(struct label *label, int flag)
{
int error;
mac_init_label(label);
MAC_CHECK(init_socket_peer_label, label, flag);
if (error) {
MAC_PERFORM(destroy_socket_label, label);
mac_destroy_label(label);
}
return (error);
}
int
mac_init_socket(struct socket *socket, int flag)
{
int error;
error = mac_init_socket_label(&socket->so_label, flag);
if (error)
return (error);
error = mac_init_socket_peer_label(&socket->so_peerlabel, flag);
if (error)
mac_destroy_socket_label(&socket->so_label);
return (error);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
Slightly change the semantics of vnode labels for MAC: rather than "refreshing" the label on the vnode before use, just get the label right from inception. For single-label file systems, set the label in the generic VFS getnewvnode() code; for multi-label file systems, leave the labeling up to the file system. With UFS1/2, this means reading the extended attribute during vfs_vget() as the inode is pulled off disk, rather than hitting the extended attributes frequently during operations later, improving performance. This also corrects sematics for shared vnode locks, which were not previously present in the system. This chances the cache coherrency properties WRT out-of-band access to label data, but in an acceptable form. With UFS1, there is a small race condition during automatic extended attribute start -- this is not present with UFS2, and occurs because EAs aren't available at vnode inception. We'll introduce a work around for this shortly. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-26 14:38:24 +00:00
void
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
mac_init_vnode_label(struct label *label)
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
{
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
mac_init_label(label);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
MAC_PERFORM(init_vnode_label, label);
Wrap maintenance of varios nmac{objectname} counters in MAC_DEBUG so we can avoid the cost of a large number of atomic operations if we're not interested in the object count statistics. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-16 14:21:38 +00:00
#ifdef MAC_DEBUG
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
atomic_add_int(&nmacvnodes, 1);
Wrap maintenance of varios nmac{objectname} counters in MAC_DEBUG so we can avoid the cost of a large number of atomic operations if we're not interested in the object count statistics. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-16 14:21:38 +00:00
#endif
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
void
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
mac_init_vnode(struct vnode *vp)
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
{
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
mac_init_vnode_label(&vp->v_label);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
void
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
mac_destroy_bpfdesc(struct bpf_d *bpf_d)
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
{
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
MAC_PERFORM(destroy_bpfdesc_label, &bpf_d->bd_label);
mac_destroy_label(&bpf_d->bd_label);
Wrap maintenance of varios nmac{objectname} counters in MAC_DEBUG so we can avoid the cost of a large number of atomic operations if we're not interested in the object count statistics. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-16 14:21:38 +00:00
#ifdef MAC_DEBUG
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
atomic_subtract_int(&nmacbpfdescs, 1);
Wrap maintenance of varios nmac{objectname} counters in MAC_DEBUG so we can avoid the cost of a large number of atomic operations if we're not interested in the object count statistics. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-16 14:21:38 +00:00
#endif
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
static void
mac_destroy_cred_label(struct label *label)
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
{
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
MAC_PERFORM(destroy_cred_label, label);
mac_destroy_label(label);
Wrap maintenance of varios nmac{objectname} counters in MAC_DEBUG so we can avoid the cost of a large number of atomic operations if we're not interested in the object count statistics. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-16 14:21:38 +00:00
#ifdef MAC_DEBUG
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
atomic_subtract_int(&nmaccreds, 1);
Wrap maintenance of varios nmac{objectname} counters in MAC_DEBUG so we can avoid the cost of a large number of atomic operations if we're not interested in the object count statistics. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-16 14:21:38 +00:00
#endif
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
void
mac_destroy_cred(struct ucred *cred)
{
mac_destroy_cred_label(&cred->cr_label);
}
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
void
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
mac_destroy_devfsdirent(struct devfs_dirent *de)
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
{
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
MAC_PERFORM(destroy_devfsdirent_label, &de->de_label);
mac_destroy_label(&de->de_label);
Wrap maintenance of varios nmac{objectname} counters in MAC_DEBUG so we can avoid the cost of a large number of atomic operations if we're not interested in the object count statistics. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-16 14:21:38 +00:00
#ifdef MAC_DEBUG
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
atomic_subtract_int(&nmacdevfsdirents, 1);
Wrap maintenance of varios nmac{objectname} counters in MAC_DEBUG so we can avoid the cost of a large number of atomic operations if we're not interested in the object count statistics. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-16 14:21:38 +00:00
#endif
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
static void
mac_destroy_ifnet_label(struct label *label)
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
{
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
MAC_PERFORM(destroy_ifnet_label, label);
mac_destroy_label(label);
Wrap maintenance of varios nmac{objectname} counters in MAC_DEBUG so we can avoid the cost of a large number of atomic operations if we're not interested in the object count statistics. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-16 14:21:38 +00:00
#ifdef MAC_DEBUG
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
atomic_subtract_int(&nmacifnets, 1);
Wrap maintenance of varios nmac{objectname} counters in MAC_DEBUG so we can avoid the cost of a large number of atomic operations if we're not interested in the object count statistics. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-16 14:21:38 +00:00
#endif
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
void
mac_destroy_ifnet(struct ifnet *ifp)
{
mac_destroy_ifnet_label(&ifp->if_label);
}
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
void
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
mac_destroy_ipq(struct ipq *ipq)
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
{
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
MAC_PERFORM(destroy_ipq_label, &ipq->ipq_label);
mac_destroy_label(&ipq->ipq_label);
Wrap maintenance of varios nmac{objectname} counters in MAC_DEBUG so we can avoid the cost of a large number of atomic operations if we're not interested in the object count statistics. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-16 14:21:38 +00:00
#ifdef MAC_DEBUG
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
atomic_subtract_int(&nmacipqs, 1);
Wrap maintenance of varios nmac{objectname} counters in MAC_DEBUG so we can avoid the cost of a large number of atomic operations if we're not interested in the object count statistics. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-16 14:21:38 +00:00
#endif
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
void
Move MAC label storage for mbufs into m_tags from the m_pkthdr structure, returning some additional room in the first mbuf in a chain, and avoiding feature-specific contents in the mbuf header. To do this: - Modify mbuf_to_label() to extract the tag, returning NULL if not found. - Introduce mac_init_mbuf_tag() which does most of the work mac_init_mbuf() used to do, except on an m_tag rather than an mbuf. - Scale back mac_init_mbuf() to perform m_tag allocation and invoke mac_init_mbuf_tag(). - Replace mac_destroy_mbuf() with mac_destroy_mbuf_tag(), since m_tag's are now GC'd deep in the m_tag/mbuf code rather than at a higher level when mbufs are directly free()'d. - Add mac_copy_mbuf_tag() to support m_copy_pkthdr() and related notions. - Generally change all references to mbuf labels so that they use mbuf_to_label() rather than &mbuf->m_pkthdr.label. This required no changes in the MAC policies (yay!). - Tweak mbuf release routines to not call mac_destroy_mbuf(), tag destruction takes care of it for us now. - Remove MAC magic from m_copy_pkthdr() and m_move_pkthdr() -- the existing m_tag support does all this for us. Note that we can no longer just zero the m_tag list on the target mbuf, rather, we have to delete the chain because m_tag's will already be hung off freshly allocated mbuf's. - Tweak m_tag copying routines so that if we're copying a MAC m_tag, we don't do a binary copy, rather, we initialize the new storage and do a deep copy of the label. - Remove use of MAC_FLAG_INITIALIZED in a few bizarre places having to do with mbuf header copies previously. - When an mbuf is copied in ip_input(), we no longer need to explicitly copy the label because it will get handled by the m_tag code now. - No longer any weird handling of MAC labels in if_loop.c during header copies. - Add MPC_LOADTIME_FLAG_LABELMBUFS flag to Biba, MLS, mac_test. In mac_test, handle the label==NULL case, since it can be dynamically loaded. In order to improve performance with this change, introduce the notion of "lazy MAC label allocation" -- only allocate m_tag storage for MAC labels if we're running with a policy that uses MAC labels on mbufs. Policies declare this intent by setting the MPC_LOADTIME_FLAG_LABELMBUFS flag in their load-time flags field during declaration. Note: this opens up the possibility of post-boot policy modules getting back NULL slot entries even though they have policy invariants of non-NULL slot entries, as the policy might have been loaded after the mbuf was allocated, leaving the mbuf without label storage. Policies that cannot handle this case must be declared as NOTLATE, or must be modified. - mac_labelmbufs holds the current cumulative status as to whether any policies require mbuf labeling or not. This is updated whenever the active policy set changes by the function mac_policy_updateflags(). The function iterates the list and checks whether any have the flag set. Write access to this variable is protected by the policy list; read access is currently not protected for performance reasons. This might change if it causes problems. - Add MAC_POLICY_LIST_ASSERT_EXCLUSIVE() to permit the flags update function to assert appropriate locks. - This makes allocation in mac_init_mbuf() conditional on the flag. Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 20:39:06 +00:00
mac_destroy_mbuf_tag(struct m_tag *tag)
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
{
Move MAC label storage for mbufs into m_tags from the m_pkthdr structure, returning some additional room in the first mbuf in a chain, and avoiding feature-specific contents in the mbuf header. To do this: - Modify mbuf_to_label() to extract the tag, returning NULL if not found. - Introduce mac_init_mbuf_tag() which does most of the work mac_init_mbuf() used to do, except on an m_tag rather than an mbuf. - Scale back mac_init_mbuf() to perform m_tag allocation and invoke mac_init_mbuf_tag(). - Replace mac_destroy_mbuf() with mac_destroy_mbuf_tag(), since m_tag's are now GC'd deep in the m_tag/mbuf code rather than at a higher level when mbufs are directly free()'d. - Add mac_copy_mbuf_tag() to support m_copy_pkthdr() and related notions. - Generally change all references to mbuf labels so that they use mbuf_to_label() rather than &mbuf->m_pkthdr.label. This required no changes in the MAC policies (yay!). - Tweak mbuf release routines to not call mac_destroy_mbuf(), tag destruction takes care of it for us now. - Remove MAC magic from m_copy_pkthdr() and m_move_pkthdr() -- the existing m_tag support does all this for us. Note that we can no longer just zero the m_tag list on the target mbuf, rather, we have to delete the chain because m_tag's will already be hung off freshly allocated mbuf's. - Tweak m_tag copying routines so that if we're copying a MAC m_tag, we don't do a binary copy, rather, we initialize the new storage and do a deep copy of the label. - Remove use of MAC_FLAG_INITIALIZED in a few bizarre places having to do with mbuf header copies previously. - When an mbuf is copied in ip_input(), we no longer need to explicitly copy the label because it will get handled by the m_tag code now. - No longer any weird handling of MAC labels in if_loop.c during header copies. - Add MPC_LOADTIME_FLAG_LABELMBUFS flag to Biba, MLS, mac_test. In mac_test, handle the label==NULL case, since it can be dynamically loaded. In order to improve performance with this change, introduce the notion of "lazy MAC label allocation" -- only allocate m_tag storage for MAC labels if we're running with a policy that uses MAC labels on mbufs. Policies declare this intent by setting the MPC_LOADTIME_FLAG_LABELMBUFS flag in their load-time flags field during declaration. Note: this opens up the possibility of post-boot policy modules getting back NULL slot entries even though they have policy invariants of non-NULL slot entries, as the policy might have been loaded after the mbuf was allocated, leaving the mbuf without label storage. Policies that cannot handle this case must be declared as NOTLATE, or must be modified. - mac_labelmbufs holds the current cumulative status as to whether any policies require mbuf labeling or not. This is updated whenever the active policy set changes by the function mac_policy_updateflags(). The function iterates the list and checks whether any have the flag set. Write access to this variable is protected by the policy list; read access is currently not protected for performance reasons. This might change if it causes problems. - Add MAC_POLICY_LIST_ASSERT_EXCLUSIVE() to permit the flags update function to assert appropriate locks. - This makes allocation in mac_init_mbuf() conditional on the flag. Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 20:39:06 +00:00
struct label *label;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
Move MAC label storage for mbufs into m_tags from the m_pkthdr structure, returning some additional room in the first mbuf in a chain, and avoiding feature-specific contents in the mbuf header. To do this: - Modify mbuf_to_label() to extract the tag, returning NULL if not found. - Introduce mac_init_mbuf_tag() which does most of the work mac_init_mbuf() used to do, except on an m_tag rather than an mbuf. - Scale back mac_init_mbuf() to perform m_tag allocation and invoke mac_init_mbuf_tag(). - Replace mac_destroy_mbuf() with mac_destroy_mbuf_tag(), since m_tag's are now GC'd deep in the m_tag/mbuf code rather than at a higher level when mbufs are directly free()'d. - Add mac_copy_mbuf_tag() to support m_copy_pkthdr() and related notions. - Generally change all references to mbuf labels so that they use mbuf_to_label() rather than &mbuf->m_pkthdr.label. This required no changes in the MAC policies (yay!). - Tweak mbuf release routines to not call mac_destroy_mbuf(), tag destruction takes care of it for us now. - Remove MAC magic from m_copy_pkthdr() and m_move_pkthdr() -- the existing m_tag support does all this for us. Note that we can no longer just zero the m_tag list on the target mbuf, rather, we have to delete the chain because m_tag's will already be hung off freshly allocated mbuf's. - Tweak m_tag copying routines so that if we're copying a MAC m_tag, we don't do a binary copy, rather, we initialize the new storage and do a deep copy of the label. - Remove use of MAC_FLAG_INITIALIZED in a few bizarre places having to do with mbuf header copies previously. - When an mbuf is copied in ip_input(), we no longer need to explicitly copy the label because it will get handled by the m_tag code now. - No longer any weird handling of MAC labels in if_loop.c during header copies. - Add MPC_LOADTIME_FLAG_LABELMBUFS flag to Biba, MLS, mac_test. In mac_test, handle the label==NULL case, since it can be dynamically loaded. In order to improve performance with this change, introduce the notion of "lazy MAC label allocation" -- only allocate m_tag storage for MAC labels if we're running with a policy that uses MAC labels on mbufs. Policies declare this intent by setting the MPC_LOADTIME_FLAG_LABELMBUFS flag in their load-time flags field during declaration. Note: this opens up the possibility of post-boot policy modules getting back NULL slot entries even though they have policy invariants of non-NULL slot entries, as the policy might have been loaded after the mbuf was allocated, leaving the mbuf without label storage. Policies that cannot handle this case must be declared as NOTLATE, or must be modified. - mac_labelmbufs holds the current cumulative status as to whether any policies require mbuf labeling or not. This is updated whenever the active policy set changes by the function mac_policy_updateflags(). The function iterates the list and checks whether any have the flag set. Write access to this variable is protected by the policy list; read access is currently not protected for performance reasons. This might change if it causes problems. - Add MAC_POLICY_LIST_ASSERT_EXCLUSIVE() to permit the flags update function to assert appropriate locks. - This makes allocation in mac_init_mbuf() conditional on the flag. Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 20:39:06 +00:00
label = (struct label *)(tag+1);
MAC_PERFORM(destroy_mbuf_label, label);
mac_destroy_label(label);
Wrap maintenance of varios nmac{objectname} counters in MAC_DEBUG so we can avoid the cost of a large number of atomic operations if we're not interested in the object count statistics. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-16 14:21:38 +00:00
#ifdef MAC_DEBUG
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
atomic_subtract_int(&nmacmbufs, 1);
Wrap maintenance of varios nmac{objectname} counters in MAC_DEBUG so we can avoid the cost of a large number of atomic operations if we're not interested in the object count statistics. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-16 14:21:38 +00:00
#endif
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
void
mac_destroy_mount(struct mount *mp)
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
{
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
MAC_PERFORM(destroy_mount_label, &mp->mnt_mntlabel);
MAC_PERFORM(destroy_mount_fs_label, &mp->mnt_fslabel);
mac_destroy_label(&mp->mnt_fslabel);
mac_destroy_label(&mp->mnt_mntlabel);
Wrap maintenance of varios nmac{objectname} counters in MAC_DEBUG so we can avoid the cost of a large number of atomic operations if we're not interested in the object count statistics. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-16 14:21:38 +00:00
#ifdef MAC_DEBUG
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
atomic_subtract_int(&nmacmounts, 1);
Wrap maintenance of varios nmac{objectname} counters in MAC_DEBUG so we can avoid the cost of a large number of atomic operations if we're not interested in the object count statistics. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-16 14:21:38 +00:00
#endif
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
static void
mac_destroy_pipe_label(struct label *label)
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
{
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
MAC_PERFORM(destroy_pipe_label, label);
mac_destroy_label(label);
Wrap maintenance of varios nmac{objectname} counters in MAC_DEBUG so we can avoid the cost of a large number of atomic operations if we're not interested in the object count statistics. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-16 14:21:38 +00:00
#ifdef MAC_DEBUG
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
atomic_subtract_int(&nmacpipes, 1);
Wrap maintenance of varios nmac{objectname} counters in MAC_DEBUG so we can avoid the cost of a large number of atomic operations if we're not interested in the object count statistics. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-16 14:21:38 +00:00
#endif
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
void
mac_destroy_pipe(struct pipe *pipe)
{
mac_destroy_pipe_label(pipe->pipe_label);
free(pipe->pipe_label, M_MACPIPELABEL);
}
Introduce p_label, extensible security label storage for the MAC framework in struct proc. While the process label is actually stored in the struct ucred pointed to by p_ucred, there is a need for transient storage that may be used when asynchronous (deferred) updates need to be performed on the "real" label for locking reasons. Unlike other label storage, this label has no locking semantics, relying on policies to provide their own protection for the label contents, meaning that a policy leaf mutex may be used, avoiding lock order issues. This permits policies that act based on historical process behavior (such as audit policies, the MAC Framework port of LOMAC, etc) can update process properties even when many existing locks are held without violating the lock order. No currently committed policies implement use of this label storage. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-11-20 15:41:25 +00:00
void
mac_destroy_proc(struct proc *p)
{
MAC_PERFORM(destroy_proc_label, &p->p_label);
mac_destroy_label(&p->p_label);
#ifdef MAC_DEBUG
atomic_subtract_int(&nmacprocs, 1);
#endif
}
Modify label allocation semantics for sockets: pass in soalloc's malloc flags so that we can call malloc with M_NOWAIT if necessary, avoiding potential sleeps while holding mutexes in the TCP syncache code. Similar to the existing support for mbuf label allocation: if we can't allocate all the necessary label store in each policy, we back out the label allocation and fail the socket creation. Sync from MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 21:23:47 +00:00
static void
mac_destroy_socket_label(struct label *label)
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
{
Modify label allocation semantics for sockets: pass in soalloc's malloc flags so that we can call malloc with M_NOWAIT if necessary, avoiding potential sleeps while holding mutexes in the TCP syncache code. Similar to the existing support for mbuf label allocation: if we can't allocate all the necessary label store in each policy, we back out the label allocation and fail the socket creation. Sync from MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 21:23:47 +00:00
MAC_PERFORM(destroy_socket_label, label);
mac_destroy_label(label);
Wrap maintenance of varios nmac{objectname} counters in MAC_DEBUG so we can avoid the cost of a large number of atomic operations if we're not interested in the object count statistics. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-16 14:21:38 +00:00
#ifdef MAC_DEBUG
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
atomic_subtract_int(&nmacsockets, 1);
Wrap maintenance of varios nmac{objectname} counters in MAC_DEBUG so we can avoid the cost of a large number of atomic operations if we're not interested in the object count statistics. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-16 14:21:38 +00:00
#endif
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
Modify label allocation semantics for sockets: pass in soalloc's malloc flags so that we can call malloc with M_NOWAIT if necessary, avoiding potential sleeps while holding mutexes in the TCP syncache code. Similar to the existing support for mbuf label allocation: if we can't allocate all the necessary label store in each policy, we back out the label allocation and fail the socket creation. Sync from MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 21:23:47 +00:00
static void
mac_destroy_socket_peer_label(struct label *label)
{
MAC_PERFORM(destroy_socket_peer_label, label);
mac_destroy_label(label);
}
void
mac_destroy_socket(struct socket *socket)
{
mac_destroy_socket_label(&socket->so_label);
mac_destroy_socket_peer_label(&socket->so_peerlabel);
}
Slightly change the semantics of vnode labels for MAC: rather than "refreshing" the label on the vnode before use, just get the label right from inception. For single-label file systems, set the label in the generic VFS getnewvnode() code; for multi-label file systems, leave the labeling up to the file system. With UFS1/2, this means reading the extended attribute during vfs_vget() as the inode is pulled off disk, rather than hitting the extended attributes frequently during operations later, improving performance. This also corrects sematics for shared vnode locks, which were not previously present in the system. This chances the cache coherrency properties WRT out-of-band access to label data, but in an acceptable form. With UFS1, there is a small race condition during automatic extended attribute start -- this is not present with UFS2, and occurs because EAs aren't available at vnode inception. We'll introduce a work around for this shortly. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-26 14:38:24 +00:00
void
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
mac_destroy_vnode_label(struct label *label)
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
{
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
MAC_PERFORM(destroy_vnode_label, label);
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
mac_destroy_label(label);
Wrap maintenance of varios nmac{objectname} counters in MAC_DEBUG so we can avoid the cost of a large number of atomic operations if we're not interested in the object count statistics. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-16 14:21:38 +00:00
#ifdef MAC_DEBUG
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
atomic_subtract_int(&nmacvnodes, 1);
Wrap maintenance of varios nmac{objectname} counters in MAC_DEBUG so we can avoid the cost of a large number of atomic operations if we're not interested in the object count statistics. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-16 14:21:38 +00:00
#endif
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
void
Rearrange object and label init/destroy functions to match the order used in mac_policy.h and elsewhere. Sort order is basically "by operation category", then "alphabetically by object". Sync to MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 17:38:45 +00:00
mac_destroy_vnode(struct vnode *vp)
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
{
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
mac_destroy_vnode_label(&vp->v_label);
}
Move MAC label storage for mbufs into m_tags from the m_pkthdr structure, returning some additional room in the first mbuf in a chain, and avoiding feature-specific contents in the mbuf header. To do this: - Modify mbuf_to_label() to extract the tag, returning NULL if not found. - Introduce mac_init_mbuf_tag() which does most of the work mac_init_mbuf() used to do, except on an m_tag rather than an mbuf. - Scale back mac_init_mbuf() to perform m_tag allocation and invoke mac_init_mbuf_tag(). - Replace mac_destroy_mbuf() with mac_destroy_mbuf_tag(), since m_tag's are now GC'd deep in the m_tag/mbuf code rather than at a higher level when mbufs are directly free()'d. - Add mac_copy_mbuf_tag() to support m_copy_pkthdr() and related notions. - Generally change all references to mbuf labels so that they use mbuf_to_label() rather than &mbuf->m_pkthdr.label. This required no changes in the MAC policies (yay!). - Tweak mbuf release routines to not call mac_destroy_mbuf(), tag destruction takes care of it for us now. - Remove MAC magic from m_copy_pkthdr() and m_move_pkthdr() -- the existing m_tag support does all this for us. Note that we can no longer just zero the m_tag list on the target mbuf, rather, we have to delete the chain because m_tag's will already be hung off freshly allocated mbuf's. - Tweak m_tag copying routines so that if we're copying a MAC m_tag, we don't do a binary copy, rather, we initialize the new storage and do a deep copy of the label. - Remove use of MAC_FLAG_INITIALIZED in a few bizarre places having to do with mbuf header copies previously. - When an mbuf is copied in ip_input(), we no longer need to explicitly copy the label because it will get handled by the m_tag code now. - No longer any weird handling of MAC labels in if_loop.c during header copies. - Add MPC_LOADTIME_FLAG_LABELMBUFS flag to Biba, MLS, mac_test. In mac_test, handle the label==NULL case, since it can be dynamically loaded. In order to improve performance with this change, introduce the notion of "lazy MAC label allocation" -- only allocate m_tag storage for MAC labels if we're running with a policy that uses MAC labels on mbufs. Policies declare this intent by setting the MPC_LOADTIME_FLAG_LABELMBUFS flag in their load-time flags field during declaration. Note: this opens up the possibility of post-boot policy modules getting back NULL slot entries even though they have policy invariants of non-NULL slot entries, as the policy might have been loaded after the mbuf was allocated, leaving the mbuf without label storage. Policies that cannot handle this case must be declared as NOTLATE, or must be modified. - mac_labelmbufs holds the current cumulative status as to whether any policies require mbuf labeling or not. This is updated whenever the active policy set changes by the function mac_policy_updateflags(). The function iterates the list and checks whether any have the flag set. Write access to this variable is protected by the policy list; read access is currently not protected for performance reasons. This might change if it causes problems. - Add MAC_POLICY_LIST_ASSERT_EXCLUSIVE() to permit the flags update function to assert appropriate locks. - This makes allocation in mac_init_mbuf() conditional on the flag. Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 20:39:06 +00:00
void
mac_copy_mbuf_tag(struct m_tag *src, struct m_tag *dest)
{
struct label *src_label, *dest_label;
src_label = (struct label *)(src+1);
dest_label = (struct label *)(dest+1);
/*
* mac_init_mbuf_tag() is called on the target tag in
* m_tag_copy(), so we don't need to call it here.
*/
MAC_PERFORM(copy_mbuf_label, src_label, dest_label);
}
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
static void
mac_copy_pipe_label(struct label *src, struct label *dest)
{
MAC_PERFORM(copy_pipe_label, src, dest);
}
Slightly change the semantics of vnode labels for MAC: rather than "refreshing" the label on the vnode before use, just get the label right from inception. For single-label file systems, set the label in the generic VFS getnewvnode() code; for multi-label file systems, leave the labeling up to the file system. With UFS1/2, this means reading the extended attribute during vfs_vget() as the inode is pulled off disk, rather than hitting the extended attributes frequently during operations later, improving performance. This also corrects sematics for shared vnode locks, which were not previously present in the system. This chances the cache coherrency properties WRT out-of-band access to label data, but in an acceptable form. With UFS1, there is a small race condition during automatic extended attribute start -- this is not present with UFS2, and occurs because EAs aren't available at vnode inception. We'll introduce a work around for this shortly. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-26 14:38:24 +00:00
void
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
mac_copy_vnode_label(struct label *src, struct label *dest)
{
MAC_PERFORM(copy_vnode_label, src, dest);
}
static int
mac_check_structmac_consistent(struct mac *mac)
{
if (mac->m_buflen > MAC_MAX_LABEL_BUF_LEN)
return (EINVAL);
return (0);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
Another big diff, little functional change: move label internalization, externalization, and cred label life cycle events to entirely above devfs and vnode events. Sync from MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:57:16 +00:00
static int
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
mac_externalize_cred_label(struct label *label, char *elements,
char *outbuf, size_t outbuflen, int flags)
Another big diff, little functional change: move label internalization, externalization, and cred label life cycle events to entirely above devfs and vnode events. Sync from MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:57:16 +00:00
{
int error;
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
MAC_EXTERNALIZE(cred_label, label, elements, outbuf, outbuflen);
Another big diff, little functional change: move label internalization, externalization, and cred label life cycle events to entirely above devfs and vnode events. Sync from MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:57:16 +00:00
return (error);
}
static int
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
mac_externalize_ifnet_label(struct label *label, char *elements,
char *outbuf, size_t outbuflen, int flags)
Another big diff, little functional change: move label internalization, externalization, and cred label life cycle events to entirely above devfs and vnode events. Sync from MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:57:16 +00:00
{
int error;
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
MAC_EXTERNALIZE(ifnet_label, label, elements, outbuf, outbuflen);
return (error);
}
static int
mac_externalize_pipe_label(struct label *label, char *elements,
char *outbuf, size_t outbuflen, int flags)
{
int error;
MAC_EXTERNALIZE(pipe_label, label, elements, outbuf, outbuflen);
return (error);
}
static int
mac_externalize_socket_label(struct label *label, char *elements,
char *outbuf, size_t outbuflen, int flags)
{
int error;
MAC_EXTERNALIZE(socket_label, label, elements, outbuf, outbuflen);
return (error);
}
static int
mac_externalize_socket_peer_label(struct label *label, char *elements,
char *outbuf, size_t outbuflen, int flags)
{
int error;
MAC_EXTERNALIZE(socket_peer_label, label, elements, outbuf, outbuflen);
return (error);
}
static int
mac_externalize_vnode_label(struct label *label, char *elements,
char *outbuf, size_t outbuflen, int flags)
{
int error;
MAC_EXTERNALIZE(vnode_label, label, elements, outbuf, outbuflen);
return (error);
}
static int
mac_internalize_cred_label(struct label *label, char *string)
{
int error;
MAC_INTERNALIZE(cred_label, label, string);
return (error);
}
static int
mac_internalize_ifnet_label(struct label *label, char *string)
{
int error;
MAC_INTERNALIZE(ifnet_label, label, string);
return (error);
}
static int
mac_internalize_pipe_label(struct label *label, char *string)
{
int error;
MAC_INTERNALIZE(pipe_label, label, string);
return (error);
}
static int
mac_internalize_socket_label(struct label *label, char *string)
{
int error;
MAC_INTERNALIZE(socket_label, label, string);
return (error);
}
static int
mac_internalize_vnode_label(struct label *label, char *string)
{
int error;
MAC_INTERNALIZE(vnode_label, label, string);
Another big diff, little functional change: move label internalization, externalization, and cred label life cycle events to entirely above devfs and vnode events. Sync from MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:57:16 +00:00
return (error);
}
/*
* Initialize MAC label for the first kernel process, from which other
* kernel processes and threads are spawned.
*/
void
mac_create_proc0(struct ucred *cred)
{
MAC_PERFORM(create_proc0, cred);
}
/*
* Initialize MAC label for the first userland process, from which other
* userland processes and threads are spawned.
*/
void
mac_create_proc1(struct ucred *cred)
{
MAC_PERFORM(create_proc1, cred);
}
void
mac_thread_userret(struct thread *td)
{
MAC_PERFORM(thread_userret, td);
}
/*
* When a new process is created, its label must be initialized. Generally,
* this involves inheritence from the parent process, modulo possible
* deltas. This function allows that processing to take place.
*/
void
mac_create_cred(struct ucred *parent_cred, struct ucred *child_cred)
{
MAC_PERFORM(create_cred, parent_cred, child_cred);
}
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
void
Remove dm_root entry from struct devfs_mount. It's never set, and is unused. Replace it with a dm_mount back-pointer to the struct mount that the devfs_mount is associated with. Export that pointer to MAC Framework entry points, where all current policies don't use the pointer. This permits the SEBSD port of SELinux's FLASK/TE to compile out-of-the-box on 5.0-CURRENT with full file system labeling support. Approved by: re (murray) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-12-09 03:44:28 +00:00
mac_update_devfsdirent(struct mount *mp, struct devfs_dirent *de,
struct vnode *vp)
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
{
Remove dm_root entry from struct devfs_mount. It's never set, and is unused. Replace it with a dm_mount back-pointer to the struct mount that the devfs_mount is associated with. Export that pointer to MAC Framework entry points, where all current policies don't use the pointer. This permits the SEBSD port of SELinux's FLASK/TE to compile out-of-the-box on 5.0-CURRENT with full file system labeling support. Approved by: re (murray) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-12-09 03:44:28 +00:00
MAC_PERFORM(update_devfsdirent, mp, de, &de->de_label, vp,
&vp->v_label);
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
}
void
Slightly change the semantics of vnode labels for MAC: rather than "refreshing" the label on the vnode before use, just get the label right from inception. For single-label file systems, set the label in the generic VFS getnewvnode() code; for multi-label file systems, leave the labeling up to the file system. With UFS1/2, this means reading the extended attribute during vfs_vget() as the inode is pulled off disk, rather than hitting the extended attributes frequently during operations later, improving performance. This also corrects sematics for shared vnode locks, which were not previously present in the system. This chances the cache coherrency properties WRT out-of-band access to label data, but in an acceptable form. With UFS1, there is a small race condition during automatic extended attribute start -- this is not present with UFS2, and occurs because EAs aren't available at vnode inception. We'll introduce a work around for this shortly. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-26 14:38:24 +00:00
mac_associate_vnode_devfs(struct mount *mp, struct devfs_dirent *de,
struct vnode *vp)
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
{
Slightly change the semantics of vnode labels for MAC: rather than "refreshing" the label on the vnode before use, just get the label right from inception. For single-label file systems, set the label in the generic VFS getnewvnode() code; for multi-label file systems, leave the labeling up to the file system. With UFS1/2, this means reading the extended attribute during vfs_vget() as the inode is pulled off disk, rather than hitting the extended attributes frequently during operations later, improving performance. This also corrects sematics for shared vnode locks, which were not previously present in the system. This chances the cache coherrency properties WRT out-of-band access to label data, but in an acceptable form. With UFS1, there is a small race condition during automatic extended attribute start -- this is not present with UFS2, and occurs because EAs aren't available at vnode inception. We'll introduce a work around for this shortly. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-26 14:38:24 +00:00
MAC_PERFORM(associate_vnode_devfs, mp, &mp->mnt_fslabel, de,
&de->de_label, vp, &vp->v_label);
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
}
Slightly change the semantics of vnode labels for MAC: rather than "refreshing" the label on the vnode before use, just get the label right from inception. For single-label file systems, set the label in the generic VFS getnewvnode() code; for multi-label file systems, leave the labeling up to the file system. With UFS1/2, this means reading the extended attribute during vfs_vget() as the inode is pulled off disk, rather than hitting the extended attributes frequently during operations later, improving performance. This also corrects sematics for shared vnode locks, which were not previously present in the system. This chances the cache coherrency properties WRT out-of-band access to label data, but in an acceptable form. With UFS1, there is a small race condition during automatic extended attribute start -- this is not present with UFS2, and occurs because EAs aren't available at vnode inception. We'll introduce a work around for this shortly. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-26 14:38:24 +00:00
int
mac_associate_vnode_extattr(struct mount *mp, struct vnode *vp)
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
{
int error;
Slightly change the semantics of vnode labels for MAC: rather than "refreshing" the label on the vnode before use, just get the label right from inception. For single-label file systems, set the label in the generic VFS getnewvnode() code; for multi-label file systems, leave the labeling up to the file system. With UFS1/2, this means reading the extended attribute during vfs_vget() as the inode is pulled off disk, rather than hitting the extended attributes frequently during operations later, improving performance. This also corrects sematics for shared vnode locks, which were not previously present in the system. This chances the cache coherrency properties WRT out-of-band access to label data, but in an acceptable form. With UFS1, there is a small race condition during automatic extended attribute start -- this is not present with UFS2, and occurs because EAs aren't available at vnode inception. We'll introduce a work around for this shortly. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-26 14:38:24 +00:00
ASSERT_VOP_LOCKED(vp, "mac_associate_vnode_extattr");
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
Slightly change the semantics of vnode labels for MAC: rather than "refreshing" the label on the vnode before use, just get the label right from inception. For single-label file systems, set the label in the generic VFS getnewvnode() code; for multi-label file systems, leave the labeling up to the file system. With UFS1/2, this means reading the extended attribute during vfs_vget() as the inode is pulled off disk, rather than hitting the extended attributes frequently during operations later, improving performance. This also corrects sematics for shared vnode locks, which were not previously present in the system. This chances the cache coherrency properties WRT out-of-band access to label data, but in an acceptable form. With UFS1, there is a small race condition during automatic extended attribute start -- this is not present with UFS2, and occurs because EAs aren't available at vnode inception. We'll introduce a work around for this shortly. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-26 14:38:24 +00:00
MAC_CHECK(associate_vnode_extattr, mp, &mp->mnt_fslabel, vp,
&vp->v_label);
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
return (error);
}
void
Slightly change the semantics of vnode labels for MAC: rather than "refreshing" the label on the vnode before use, just get the label right from inception. For single-label file systems, set the label in the generic VFS getnewvnode() code; for multi-label file systems, leave the labeling up to the file system. With UFS1/2, this means reading the extended attribute during vfs_vget() as the inode is pulled off disk, rather than hitting the extended attributes frequently during operations later, improving performance. This also corrects sematics for shared vnode locks, which were not previously present in the system. This chances the cache coherrency properties WRT out-of-band access to label data, but in an acceptable form. With UFS1, there is a small race condition during automatic extended attribute start -- this is not present with UFS2, and occurs because EAs aren't available at vnode inception. We'll introduce a work around for this shortly. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-26 14:38:24 +00:00
mac_associate_vnode_singlelabel(struct mount *mp, struct vnode *vp)
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
{
Slightly change the semantics of vnode labels for MAC: rather than "refreshing" the label on the vnode before use, just get the label right from inception. For single-label file systems, set the label in the generic VFS getnewvnode() code; for multi-label file systems, leave the labeling up to the file system. With UFS1/2, this means reading the extended attribute during vfs_vget() as the inode is pulled off disk, rather than hitting the extended attributes frequently during operations later, improving performance. This also corrects sematics for shared vnode locks, which were not previously present in the system. This chances the cache coherrency properties WRT out-of-band access to label data, but in an acceptable form. With UFS1, there is a small race condition during automatic extended attribute start -- this is not present with UFS2, and occurs because EAs aren't available at vnode inception. We'll introduce a work around for this shortly. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-26 14:38:24 +00:00
MAC_PERFORM(associate_vnode_singlelabel, mp, &mp->mnt_fslabel, vp,
&vp->v_label);
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
}
int
Slightly change the semantics of vnode labels for MAC: rather than "refreshing" the label on the vnode before use, just get the label right from inception. For single-label file systems, set the label in the generic VFS getnewvnode() code; for multi-label file systems, leave the labeling up to the file system. With UFS1/2, this means reading the extended attribute during vfs_vget() as the inode is pulled off disk, rather than hitting the extended attributes frequently during operations later, improving performance. This also corrects sematics for shared vnode locks, which were not previously present in the system. This chances the cache coherrency properties WRT out-of-band access to label data, but in an acceptable form. With UFS1, there is a small race condition during automatic extended attribute start -- this is not present with UFS2, and occurs because EAs aren't available at vnode inception. We'll introduce a work around for this shortly. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-26 14:38:24 +00:00
mac_create_vnode_extattr(struct ucred *cred, struct mount *mp,
struct vnode *dvp, struct vnode *vp, struct componentname *cnp)
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
{
Slightly change the semantics of vnode labels for MAC: rather than "refreshing" the label on the vnode before use, just get the label right from inception. For single-label file systems, set the label in the generic VFS getnewvnode() code; for multi-label file systems, leave the labeling up to the file system. With UFS1/2, this means reading the extended attribute during vfs_vget() as the inode is pulled off disk, rather than hitting the extended attributes frequently during operations later, improving performance. This also corrects sematics for shared vnode locks, which were not previously present in the system. This chances the cache coherrency properties WRT out-of-band access to label data, but in an acceptable form. With UFS1, there is a small race condition during automatic extended attribute start -- this is not present with UFS2, and occurs because EAs aren't available at vnode inception. We'll introduce a work around for this shortly. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-26 14:38:24 +00:00
int error;
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
Slightly change the semantics of vnode labels for MAC: rather than "refreshing" the label on the vnode before use, just get the label right from inception. For single-label file systems, set the label in the generic VFS getnewvnode() code; for multi-label file systems, leave the labeling up to the file system. With UFS1/2, this means reading the extended attribute during vfs_vget() as the inode is pulled off disk, rather than hitting the extended attributes frequently during operations later, improving performance. This also corrects sematics for shared vnode locks, which were not previously present in the system. This chances the cache coherrency properties WRT out-of-band access to label data, but in an acceptable form. With UFS1, there is a small race condition during automatic extended attribute start -- this is not present with UFS2, and occurs because EAs aren't available at vnode inception. We'll introduce a work around for this shortly. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-26 14:38:24 +00:00
ASSERT_VOP_LOCKED(dvp, "mac_create_vnode_extattr");
ASSERT_VOP_LOCKED(vp, "mac_create_vnode_extattr");
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
Slightly change the semantics of vnode labels for MAC: rather than "refreshing" the label on the vnode before use, just get the label right from inception. For single-label file systems, set the label in the generic VFS getnewvnode() code; for multi-label file systems, leave the labeling up to the file system. With UFS1/2, this means reading the extended attribute during vfs_vget() as the inode is pulled off disk, rather than hitting the extended attributes frequently during operations later, improving performance. This also corrects sematics for shared vnode locks, which were not previously present in the system. This chances the cache coherrency properties WRT out-of-band access to label data, but in an acceptable form. With UFS1, there is a small race condition during automatic extended attribute start -- this is not present with UFS2, and occurs because EAs aren't available at vnode inception. We'll introduce a work around for this shortly. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-26 14:38:24 +00:00
error = VOP_OPENEXTATTR(vp, cred, curthread);
if (error == EOPNOTSUPP) {
/* XXX: Optionally abort if transactions not supported. */
if (ea_warn_once == 0) {
printf("Warning: transactions not supported "
"in EA write.\n");
ea_warn_once = 1;
}
} else if (error)
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
return (error);
Slightly change the semantics of vnode labels for MAC: rather than "refreshing" the label on the vnode before use, just get the label right from inception. For single-label file systems, set the label in the generic VFS getnewvnode() code; for multi-label file systems, leave the labeling up to the file system. With UFS1/2, this means reading the extended attribute during vfs_vget() as the inode is pulled off disk, rather than hitting the extended attributes frequently during operations later, improving performance. This also corrects sematics for shared vnode locks, which were not previously present in the system. This chances the cache coherrency properties WRT out-of-band access to label data, but in an acceptable form. With UFS1, there is a small race condition during automatic extended attribute start -- this is not present with UFS2, and occurs because EAs aren't available at vnode inception. We'll introduce a work around for this shortly. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-26 14:38:24 +00:00
MAC_CHECK(create_vnode_extattr, cred, mp, &mp->mnt_fslabel,
dvp, &dvp->v_label, vp, &vp->v_label, cnp);
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
Slightly change the semantics of vnode labels for MAC: rather than "refreshing" the label on the vnode before use, just get the label right from inception. For single-label file systems, set the label in the generic VFS getnewvnode() code; for multi-label file systems, leave the labeling up to the file system. With UFS1/2, this means reading the extended attribute during vfs_vget() as the inode is pulled off disk, rather than hitting the extended attributes frequently during operations later, improving performance. This also corrects sematics for shared vnode locks, which were not previously present in the system. This chances the cache coherrency properties WRT out-of-band access to label data, but in an acceptable form. With UFS1, there is a small race condition during automatic extended attribute start -- this is not present with UFS2, and occurs because EAs aren't available at vnode inception. We'll introduce a work around for this shortly. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-26 14:38:24 +00:00
if (error) {
VOP_CLOSEEXTATTR(vp, 0, NOCRED, curthread);
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
return (error);
}
Slightly change the semantics of vnode labels for MAC: rather than "refreshing" the label on the vnode before use, just get the label right from inception. For single-label file systems, set the label in the generic VFS getnewvnode() code; for multi-label file systems, leave the labeling up to the file system. With UFS1/2, this means reading the extended attribute during vfs_vget() as the inode is pulled off disk, rather than hitting the extended attributes frequently during operations later, improving performance. This also corrects sematics for shared vnode locks, which were not previously present in the system. This chances the cache coherrency properties WRT out-of-band access to label data, but in an acceptable form. With UFS1, there is a small race condition during automatic extended attribute start -- this is not present with UFS2, and occurs because EAs aren't available at vnode inception. We'll introduce a work around for this shortly. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-26 14:38:24 +00:00
error = VOP_CLOSEEXTATTR(vp, 1, NOCRED, curthread);
if (error == EOPNOTSUPP)
error = 0; /* XXX */
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
return (error);
}
static int
Slightly change the semantics of vnode labels for MAC: rather than "refreshing" the label on the vnode before use, just get the label right from inception. For single-label file systems, set the label in the generic VFS getnewvnode() code; for multi-label file systems, leave the labeling up to the file system. With UFS1/2, this means reading the extended attribute during vfs_vget() as the inode is pulled off disk, rather than hitting the extended attributes frequently during operations later, improving performance. This also corrects sematics for shared vnode locks, which were not previously present in the system. This chances the cache coherrency properties WRT out-of-band access to label data, but in an acceptable form. With UFS1, there is a small race condition during automatic extended attribute start -- this is not present with UFS2, and occurs because EAs aren't available at vnode inception. We'll introduce a work around for this shortly. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-26 14:38:24 +00:00
mac_setlabel_vnode_extattr(struct ucred *cred, struct vnode *vp,
struct label *intlabel)
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
{
int error;
Slightly change the semantics of vnode labels for MAC: rather than "refreshing" the label on the vnode before use, just get the label right from inception. For single-label file systems, set the label in the generic VFS getnewvnode() code; for multi-label file systems, leave the labeling up to the file system. With UFS1/2, this means reading the extended attribute during vfs_vget() as the inode is pulled off disk, rather than hitting the extended attributes frequently during operations later, improving performance. This also corrects sematics for shared vnode locks, which were not previously present in the system. This chances the cache coherrency properties WRT out-of-band access to label data, but in an acceptable form. With UFS1, there is a small race condition during automatic extended attribute start -- this is not present with UFS2, and occurs because EAs aren't available at vnode inception. We'll introduce a work around for this shortly. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-26 14:38:24 +00:00
ASSERT_VOP_LOCKED(vp, "mac_setlabel_vnode_extattr");
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
Slightly change the semantics of vnode labels for MAC: rather than "refreshing" the label on the vnode before use, just get the label right from inception. For single-label file systems, set the label in the generic VFS getnewvnode() code; for multi-label file systems, leave the labeling up to the file system. With UFS1/2, this means reading the extended attribute during vfs_vget() as the inode is pulled off disk, rather than hitting the extended attributes frequently during operations later, improving performance. This also corrects sematics for shared vnode locks, which were not previously present in the system. This chances the cache coherrency properties WRT out-of-band access to label data, but in an acceptable form. With UFS1, there is a small race condition during automatic extended attribute start -- this is not present with UFS2, and occurs because EAs aren't available at vnode inception. We'll introduce a work around for this shortly. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-26 14:38:24 +00:00
error = VOP_OPENEXTATTR(vp, cred, curthread);
if (error == EOPNOTSUPP) {
/* XXX: Optionally abort if transactions not supported. */
if (ea_warn_once == 0) {
printf("Warning: transactions not supported "
"in EA write.\n");
ea_warn_once = 1;
}
} else if (error)
return (error);
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
Slightly change the semantics of vnode labels for MAC: rather than "refreshing" the label on the vnode before use, just get the label right from inception. For single-label file systems, set the label in the generic VFS getnewvnode() code; for multi-label file systems, leave the labeling up to the file system. With UFS1/2, this means reading the extended attribute during vfs_vget() as the inode is pulled off disk, rather than hitting the extended attributes frequently during operations later, improving performance. This also corrects sematics for shared vnode locks, which were not previously present in the system. This chances the cache coherrency properties WRT out-of-band access to label data, but in an acceptable form. With UFS1, there is a small race condition during automatic extended attribute start -- this is not present with UFS2, and occurs because EAs aren't available at vnode inception. We'll introduce a work around for this shortly. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-26 14:38:24 +00:00
MAC_CHECK(setlabel_vnode_extattr, cred, vp, &vp->v_label, intlabel);
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
Slightly change the semantics of vnode labels for MAC: rather than "refreshing" the label on the vnode before use, just get the label right from inception. For single-label file systems, set the label in the generic VFS getnewvnode() code; for multi-label file systems, leave the labeling up to the file system. With UFS1/2, this means reading the extended attribute during vfs_vget() as the inode is pulled off disk, rather than hitting the extended attributes frequently during operations later, improving performance. This also corrects sematics for shared vnode locks, which were not previously present in the system. This chances the cache coherrency properties WRT out-of-band access to label data, but in an acceptable form. With UFS1, there is a small race condition during automatic extended attribute start -- this is not present with UFS2, and occurs because EAs aren't available at vnode inception. We'll introduce a work around for this shortly. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-26 14:38:24 +00:00
if (error) {
VOP_CLOSEEXTATTR(vp, 0, NOCRED, curthread);
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
return (error);
}
Slightly change the semantics of vnode labels for MAC: rather than "refreshing" the label on the vnode before use, just get the label right from inception. For single-label file systems, set the label in the generic VFS getnewvnode() code; for multi-label file systems, leave the labeling up to the file system. With UFS1/2, this means reading the extended attribute during vfs_vget() as the inode is pulled off disk, rather than hitting the extended attributes frequently during operations later, improving performance. This also corrects sematics for shared vnode locks, which were not previously present in the system. This chances the cache coherrency properties WRT out-of-band access to label data, but in an acceptable form. With UFS1, there is a small race condition during automatic extended attribute start -- this is not present with UFS2, and occurs because EAs aren't available at vnode inception. We'll introduce a work around for this shortly. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-26 14:38:24 +00:00
error = VOP_CLOSEEXTATTR(vp, 1, NOCRED, curthread);
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
Slightly change the semantics of vnode labels for MAC: rather than "refreshing" the label on the vnode before use, just get the label right from inception. For single-label file systems, set the label in the generic VFS getnewvnode() code; for multi-label file systems, leave the labeling up to the file system. With UFS1/2, this means reading the extended attribute during vfs_vget() as the inode is pulled off disk, rather than hitting the extended attributes frequently during operations later, improving performance. This also corrects sematics for shared vnode locks, which were not previously present in the system. This chances the cache coherrency properties WRT out-of-band access to label data, but in an acceptable form. With UFS1, there is a small race condition during automatic extended attribute start -- this is not present with UFS2, and occurs because EAs aren't available at vnode inception. We'll introduce a work around for this shortly. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-26 14:38:24 +00:00
if (error == EOPNOTSUPP)
error = 0; /* XXX */
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
Slightly change the semantics of vnode labels for MAC: rather than "refreshing" the label on the vnode before use, just get the label right from inception. For single-label file systems, set the label in the generic VFS getnewvnode() code; for multi-label file systems, leave the labeling up to the file system. With UFS1/2, this means reading the extended attribute during vfs_vget() as the inode is pulled off disk, rather than hitting the extended attributes frequently during operations later, improving performance. This also corrects sematics for shared vnode locks, which were not previously present in the system. This chances the cache coherrency properties WRT out-of-band access to label data, but in an acceptable form. With UFS1, there is a small race condition during automatic extended attribute start -- this is not present with UFS2, and occurs because EAs aren't available at vnode inception. We'll introduce a work around for this shortly. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-26 14:38:24 +00:00
return (error);
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
}
Bring in two sets of changes: (1) Permit userland applications to request a change of label atomic with an execve() via mac_execve(). This is required for the SEBSD port of SELinux/FLASK. Attempts to invoke this without MAC compiled in result in ENOSYS, as with all other MAC system calls. Complexity, if desired, is present in policy modules, rather than the framework. (2) Permit policies to have access to both the label of the vnode being executed as well as the interpreter if it's a shell script or related UNIX nonsense. Because we can't hold both vnode locks at the same time, cache the interpreter label. SEBSD relies on this because it supports secure transitioning via shell script executables. Other policies might want to take both labels into account during an integrity or confidentiality decision at execve()-time. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-11-05 17:51:56 +00:00
int
mac_execve_enter(struct image_params *imgp, struct mac *mac_p,
struct label *execlabelstorage)
{
struct mac mac;
char *buffer;
int error;
if (mac_p == NULL)
return (0);
error = copyin(mac_p, &mac, sizeof(mac));
if (error)
return (error);
error = mac_check_structmac_consistent(&mac);
if (error)
return (error);
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
buffer = malloc(mac.m_buflen, M_MACTEMP, M_WAITOK);
Bring in two sets of changes: (1) Permit userland applications to request a change of label atomic with an execve() via mac_execve(). This is required for the SEBSD port of SELinux/FLASK. Attempts to invoke this without MAC compiled in result in ENOSYS, as with all other MAC system calls. Complexity, if desired, is present in policy modules, rather than the framework. (2) Permit policies to have access to both the label of the vnode being executed as well as the interpreter if it's a shell script or related UNIX nonsense. Because we can't hold both vnode locks at the same time, cache the interpreter label. SEBSD relies on this because it supports secure transitioning via shell script executables. Other policies might want to take both labels into account during an integrity or confidentiality decision at execve()-time. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-11-05 17:51:56 +00:00
error = copyinstr(mac.m_string, buffer, mac.m_buflen, NULL);
if (error) {
free(buffer, M_MACTEMP);
return (error);
}
mac_init_cred_label(execlabelstorage);
error = mac_internalize_cred_label(execlabelstorage, buffer);
free(buffer, M_MACTEMP);
if (error) {
mac_destroy_cred_label(execlabelstorage);
return (error);
}
imgp->execlabel = execlabelstorage;
return (0);
}
void
mac_execve_exit(struct image_params *imgp)
{
if (imgp->execlabel != NULL)
mac_destroy_cred_label(imgp->execlabel);
}
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
void
Bring in two sets of changes: (1) Permit userland applications to request a change of label atomic with an execve() via mac_execve(). This is required for the SEBSD port of SELinux/FLASK. Attempts to invoke this without MAC compiled in result in ENOSYS, as with all other MAC system calls. Complexity, if desired, is present in policy modules, rather than the framework. (2) Permit policies to have access to both the label of the vnode being executed as well as the interpreter if it's a shell script or related UNIX nonsense. Because we can't hold both vnode locks at the same time, cache the interpreter label. SEBSD relies on this because it supports secure transitioning via shell script executables. Other policies might want to take both labels into account during an integrity or confidentiality decision at execve()-time. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-11-05 17:51:56 +00:00
mac_execve_transition(struct ucred *old, struct ucred *new, struct vnode *vp,
struct label *interpvnodelabel, struct image_params *imgp)
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
{
ASSERT_VOP_LOCKED(vp, "mac_execve_transition");
Assert that appropriate vnodes are locked in mac_execve_will_transition(). Allow transitioning to be twiddled off using the process and fs enforcement flags, although at some point this should probably be its own flag. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-11-05 15:11:33 +00:00
if (!mac_enforce_process && !mac_enforce_fs)
return;
Bring in two sets of changes: (1) Permit userland applications to request a change of label atomic with an execve() via mac_execve(). This is required for the SEBSD port of SELinux/FLASK. Attempts to invoke this without MAC compiled in result in ENOSYS, as with all other MAC system calls. Complexity, if desired, is present in policy modules, rather than the framework. (2) Permit policies to have access to both the label of the vnode being executed as well as the interpreter if it's a shell script or related UNIX nonsense. Because we can't hold both vnode locks at the same time, cache the interpreter label. SEBSD relies on this because it supports secure transitioning via shell script executables. Other policies might want to take both labels into account during an integrity or confidentiality decision at execve()-time. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-11-05 17:51:56 +00:00
MAC_PERFORM(execve_transition, old, new, vp, &vp->v_label,
Add an explicit execlabel argument to exec-related MAC policy entry points, rather than relying on policies to grub around in the image activator instance structure. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-11-08 18:04:00 +00:00
interpvnodelabel, imgp, imgp->execlabel);
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
}
int
Bring in two sets of changes: (1) Permit userland applications to request a change of label atomic with an execve() via mac_execve(). This is required for the SEBSD port of SELinux/FLASK. Attempts to invoke this without MAC compiled in result in ENOSYS, as with all other MAC system calls. Complexity, if desired, is present in policy modules, rather than the framework. (2) Permit policies to have access to both the label of the vnode being executed as well as the interpreter if it's a shell script or related UNIX nonsense. Because we can't hold both vnode locks at the same time, cache the interpreter label. SEBSD relies on this because it supports secure transitioning via shell script executables. Other policies might want to take both labels into account during an integrity or confidentiality decision at execve()-time. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-11-05 17:51:56 +00:00
mac_execve_will_transition(struct ucred *old, struct vnode *vp,
struct label *interpvnodelabel, struct image_params *imgp)
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
{
Slightly change the semantics of vnode labels for MAC: rather than "refreshing" the label on the vnode before use, just get the label right from inception. For single-label file systems, set the label in the generic VFS getnewvnode() code; for multi-label file systems, leave the labeling up to the file system. With UFS1/2, this means reading the extended attribute during vfs_vget() as the inode is pulled off disk, rather than hitting the extended attributes frequently during operations later, improving performance. This also corrects sematics for shared vnode locks, which were not previously present in the system. This chances the cache coherrency properties WRT out-of-band access to label data, but in an acceptable form. With UFS1, there is a small race condition during automatic extended attribute start -- this is not present with UFS2, and occurs because EAs aren't available at vnode inception. We'll introduce a work around for this shortly. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-26 14:38:24 +00:00
int result;
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
Assert that appropriate vnodes are locked in mac_execve_will_transition(). Allow transitioning to be twiddled off using the process and fs enforcement flags, although at some point this should probably be its own flag. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-11-05 15:11:33 +00:00
ASSERT_VOP_LOCKED(vp, "mac_execve_will_transition");
if (!mac_enforce_process && !mac_enforce_fs)
return (0);
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
result = 0;
Bring in two sets of changes: (1) Permit userland applications to request a change of label atomic with an execve() via mac_execve(). This is required for the SEBSD port of SELinux/FLASK. Attempts to invoke this without MAC compiled in result in ENOSYS, as with all other MAC system calls. Complexity, if desired, is present in policy modules, rather than the framework. (2) Permit policies to have access to both the label of the vnode being executed as well as the interpreter if it's a shell script or related UNIX nonsense. Because we can't hold both vnode locks at the same time, cache the interpreter label. SEBSD relies on this because it supports secure transitioning via shell script executables. Other policies might want to take both labels into account during an integrity or confidentiality decision at execve()-time. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-11-05 17:51:56 +00:00
MAC_BOOLEAN(execve_will_transition, ||, old, vp, &vp->v_label,
Add an explicit execlabel argument to exec-related MAC policy entry points, rather than relying on policies to grub around in the image activator instance structure. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-11-08 18:04:00 +00:00
interpvnodelabel, imgp, imgp->execlabel);
Move all object label init/destroy routines to the head of the entry points to better match the entry point ordering in mac_policy.h. Big diff, no functional change; merge from the MAC tree. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 16:54:59 +00:00
return (result);
}
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
int
While 'mode_t' seemed like a good idea for the access mode argument for MAC access() and open() checks, the argument actually has an int type where it becomes available. Switch to using 'int' for the mode argument throughout the MAC Framework and policy modules. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-30 17:56:57 +00:00
mac_check_vnode_access(struct ucred *cred, struct vnode *vp, int acc_mode)
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
{
int error;
ASSERT_VOP_LOCKED(vp, "mac_check_vnode_access");
if (!mac_enforce_fs)
return (0);
While 'mode_t' seemed like a good idea for the access mode argument for MAC access() and open() checks, the argument actually has an int type where it becomes available. Switch to using 'int' for the mode argument throughout the MAC Framework and policy modules. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-30 17:56:57 +00:00
MAC_CHECK(check_vnode_access, cred, vp, &vp->v_label, acc_mode);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
return (error);
}
int
mac_check_vnode_chdir(struct ucred *cred, struct vnode *dvp)
{
int error;
ASSERT_VOP_LOCKED(dvp, "mac_check_vnode_chdir");
if (!mac_enforce_fs)
return (0);
MAC_CHECK(check_vnode_chdir, cred, dvp, &dvp->v_label);
return (error);
}
int
mac_check_vnode_chroot(struct ucred *cred, struct vnode *dvp)
{
int error;
ASSERT_VOP_LOCKED(dvp, "mac_check_vnode_chroot");
if (!mac_enforce_fs)
return (0);
MAC_CHECK(check_vnode_chroot, cred, dvp, &dvp->v_label);
return (error);
}
int
mac_check_vnode_create(struct ucred *cred, struct vnode *dvp,
struct componentname *cnp, struct vattr *vap)
{
int error;
ASSERT_VOP_LOCKED(dvp, "mac_check_vnode_create");
if (!mac_enforce_fs)
return (0);
MAC_CHECK(check_vnode_create, cred, dvp, &dvp->v_label, cnp, vap);
return (error);
}
int
mac_check_vnode_delete(struct ucred *cred, struct vnode *dvp, struct vnode *vp,
struct componentname *cnp)
{
int error;
ASSERT_VOP_LOCKED(dvp, "mac_check_vnode_delete");
ASSERT_VOP_LOCKED(vp, "mac_check_vnode_delete");
if (!mac_enforce_fs)
return (0);
MAC_CHECK(check_vnode_delete, cred, dvp, &dvp->v_label, vp,
&vp->v_label, cnp);
return (error);
}
int
mac_check_vnode_deleteacl(struct ucred *cred, struct vnode *vp,
acl_type_t type)
{
int error;
ASSERT_VOP_LOCKED(vp, "mac_check_vnode_deleteacl");
if (!mac_enforce_fs)
return (0);
MAC_CHECK(check_vnode_deleteacl, cred, vp, &vp->v_label, type);
return (error);
}
int
Bring in two sets of changes: (1) Permit userland applications to request a change of label atomic with an execve() via mac_execve(). This is required for the SEBSD port of SELinux/FLASK. Attempts to invoke this without MAC compiled in result in ENOSYS, as with all other MAC system calls. Complexity, if desired, is present in policy modules, rather than the framework. (2) Permit policies to have access to both the label of the vnode being executed as well as the interpreter if it's a shell script or related UNIX nonsense. Because we can't hold both vnode locks at the same time, cache the interpreter label. SEBSD relies on this because it supports secure transitioning via shell script executables. Other policies might want to take both labels into account during an integrity or confidentiality decision at execve()-time. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-11-05 17:51:56 +00:00
mac_check_vnode_exec(struct ucred *cred, struct vnode *vp,
struct image_params *imgp)
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
{
int error;
Add a missing vnode assertion for the exec() check. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-19 15:28:39 +00:00
ASSERT_VOP_LOCKED(vp, "mac_check_vnode_exec");
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
if (!mac_enforce_process && !mac_enforce_fs)
return (0);
Add an explicit execlabel argument to exec-related MAC policy entry points, rather than relying on policies to grub around in the image activator instance structure. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-11-08 18:04:00 +00:00
MAC_CHECK(check_vnode_exec, cred, vp, &vp->v_label, imgp,
imgp->execlabel);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
return (error);
}
int
mac_check_vnode_getacl(struct ucred *cred, struct vnode *vp, acl_type_t type)
{
int error;
ASSERT_VOP_LOCKED(vp, "mac_check_vnode_getacl");
if (!mac_enforce_fs)
return (0);
MAC_CHECK(check_vnode_getacl, cred, vp, &vp->v_label, type);
return (error);
}
int
mac_check_vnode_getextattr(struct ucred *cred, struct vnode *vp,
int attrnamespace, const char *name, struct uio *uio)
{
int error;
ASSERT_VOP_LOCKED(vp, "mac_check_vnode_getextattr");
if (!mac_enforce_fs)
return (0);
MAC_CHECK(check_vnode_getextattr, cred, vp, &vp->v_label,
attrnamespace, name, uio);
return (error);
}
Merge support for mac_check_vnode_link(), a MAC framework/policy entry point that instruments the creation of hard links. Policy implementations to follow. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 18:11:36 +00:00
int
mac_check_vnode_link(struct ucred *cred, struct vnode *dvp,
struct vnode *vp, struct componentname *cnp)
{
int error;
ASSERT_VOP_LOCKED(dvp, "mac_check_vnode_link");
ASSERT_VOP_LOCKED(vp, "mac_check_vnode_link");
if (!mac_enforce_fs)
return (0);
MAC_CHECK(check_vnode_link, cred, dvp, &dvp->v_label, vp,
&vp->v_label, cnp);
return (error);
}
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
int
mac_check_vnode_lookup(struct ucred *cred, struct vnode *dvp,
struct componentname *cnp)
{
int error;
ASSERT_VOP_LOCKED(dvp, "mac_check_vnode_lookup");
if (!mac_enforce_fs)
return (0);
MAC_CHECK(check_vnode_lookup, cred, dvp, &dvp->v_label, cnp);
return (error);
}
Sync from MAC tree: break out the single mmap entry point into seperate entry points for each occasion: mac_check_vnode_mmap() Check at initial mapping mac_check_vnode_mprotect() Check at mapping protection change mac_check_vnode_mmap_downgrade() Determine if a mapping downgrade should take place following subject relabel. Implement mmap() and mprotect() entry points for labeled vnode policies. These entry points are currently not hooked up to the VM system in the base tree. These changes improve the consistency of the access control interface and offer more flexibility regarding limiting access to vnode mmaping. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-06 02:46:26 +00:00
int
mac_check_vnode_mmap(struct ucred *cred, struct vnode *vp, int prot)
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
{
Sync from MAC tree: break out the single mmap entry point into seperate entry points for each occasion: mac_check_vnode_mmap() Check at initial mapping mac_check_vnode_mprotect() Check at mapping protection change mac_check_vnode_mmap_downgrade() Determine if a mapping downgrade should take place following subject relabel. Implement mmap() and mprotect() entry points for labeled vnode policies. These entry points are currently not hooked up to the VM system in the base tree. These changes improve the consistency of the access control interface and offer more flexibility regarding limiting access to vnode mmaping. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-06 02:46:26 +00:00
int error;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
Sync from MAC tree: break out the single mmap entry point into seperate entry points for each occasion: mac_check_vnode_mmap() Check at initial mapping mac_check_vnode_mprotect() Check at mapping protection change mac_check_vnode_mmap_downgrade() Determine if a mapping downgrade should take place following subject relabel. Implement mmap() and mprotect() entry points for labeled vnode policies. These entry points are currently not hooked up to the VM system in the base tree. These changes improve the consistency of the access control interface and offer more flexibility regarding limiting access to vnode mmaping. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-06 02:46:26 +00:00
ASSERT_VOP_LOCKED(vp, "mac_check_vnode_mmap");
Add a toggle to disable VM enforcement. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-09-18 02:02:08 +00:00
Sync from MAC tree: break out the single mmap entry point into seperate entry points for each occasion: mac_check_vnode_mmap() Check at initial mapping mac_check_vnode_mprotect() Check at mapping protection change mac_check_vnode_mmap_downgrade() Determine if a mapping downgrade should take place following subject relabel. Implement mmap() and mprotect() entry points for labeled vnode policies. These entry points are currently not hooked up to the VM system in the base tree. These changes improve the consistency of the access control interface and offer more flexibility regarding limiting access to vnode mmaping. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-06 02:46:26 +00:00
if (!mac_enforce_fs || !mac_enforce_vm)
return (0);
MAC_CHECK(check_vnode_mmap, cred, vp, &vp->v_label, prot);
return (error);
}
void
mac_check_vnode_mmap_downgrade(struct ucred *cred, struct vnode *vp, int *prot)
{
int result = *prot;
ASSERT_VOP_LOCKED(vp, "mac_check_vnode_mmap_downgrade");
if (!mac_enforce_fs || !mac_enforce_vm)
return;
MAC_PERFORM(check_vnode_mmap_downgrade, cred, vp, &vp->v_label,
&result);
*prot = result;
}
int
mac_check_vnode_mprotect(struct ucred *cred, struct vnode *vp, int prot)
{
int error;
ASSERT_VOP_LOCKED(vp, "mac_check_vnode_mprotect");
if (!mac_enforce_fs || !mac_enforce_vm)
return (0);
MAC_CHECK(check_vnode_mprotect, cred, vp, &vp->v_label, prot);
return (error);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
int
While 'mode_t' seemed like a good idea for the access mode argument for MAC access() and open() checks, the argument actually has an int type where it becomes available. Switch to using 'int' for the mode argument throughout the MAC Framework and policy modules. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-30 17:56:57 +00:00
mac_check_vnode_open(struct ucred *cred, struct vnode *vp, int acc_mode)
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
{
int error;
Break out mac_check_vnode_op() into three seperate checks: mac_check_vnode_poll(), mac_check_vnode_read(), mac_check_vnode_write(). This improves the consistency with other existing vnode checks, and allows policies to avoid implementing switch statements to determine what operations they do and do not want to authorize. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-19 16:43:25 +00:00
ASSERT_VOP_LOCKED(vp, "mac_check_vnode_open");
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
if (!mac_enforce_fs)
return (0);
Break out mac_check_vnode_op() into three seperate checks: mac_check_vnode_poll(), mac_check_vnode_read(), mac_check_vnode_write(). This improves the consistency with other existing vnode checks, and allows policies to avoid implementing switch statements to determine what operations they do and do not want to authorize. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-19 16:43:25 +00:00
MAC_CHECK(check_vnode_open, cred, vp, &vp->v_label, acc_mode);
return (error);
}
int
Pass active_cred and file_cred into the MAC framework explicitly for mac_check_vnode_{poll,read,stat,write}(). Pass in fp->f_cred when calling these checks with a struct file available. Otherwise, pass NOCRED. All currently MAC policies use active_cred, but could now offer the cached credential semantic used for the base system security model. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-19 19:04:53 +00:00
mac_check_vnode_poll(struct ucred *active_cred, struct ucred *file_cred,
struct vnode *vp)
Break out mac_check_vnode_op() into three seperate checks: mac_check_vnode_poll(), mac_check_vnode_read(), mac_check_vnode_write(). This improves the consistency with other existing vnode checks, and allows policies to avoid implementing switch statements to determine what operations they do and do not want to authorize. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-19 16:43:25 +00:00
{
int error;
ASSERT_VOP_LOCKED(vp, "mac_check_vnode_poll");
if (!mac_enforce_fs)
return (0);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
Pass active_cred and file_cred into the MAC framework explicitly for mac_check_vnode_{poll,read,stat,write}(). Pass in fp->f_cred when calling these checks with a struct file available. Otherwise, pass NOCRED. All currently MAC policies use active_cred, but could now offer the cached credential semantic used for the base system security model. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-19 19:04:53 +00:00
MAC_CHECK(check_vnode_poll, active_cred, file_cred, vp,
&vp->v_label);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
return (error);
}
int
Pass active_cred and file_cred into the MAC framework explicitly for mac_check_vnode_{poll,read,stat,write}(). Pass in fp->f_cred when calling these checks with a struct file available. Otherwise, pass NOCRED. All currently MAC policies use active_cred, but could now offer the cached credential semantic used for the base system security model. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-19 19:04:53 +00:00
mac_check_vnode_read(struct ucred *active_cred, struct ucred *file_cred,
struct vnode *vp)
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
{
int error;
Break out mac_check_vnode_op() into three seperate checks: mac_check_vnode_poll(), mac_check_vnode_read(), mac_check_vnode_write(). This improves the consistency with other existing vnode checks, and allows policies to avoid implementing switch statements to determine what operations they do and do not want to authorize. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-19 16:43:25 +00:00
ASSERT_VOP_LOCKED(vp, "mac_check_vnode_read");
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
if (!mac_enforce_fs)
return (0);
Pass active_cred and file_cred into the MAC framework explicitly for mac_check_vnode_{poll,read,stat,write}(). Pass in fp->f_cred when calling these checks with a struct file available. Otherwise, pass NOCRED. All currently MAC policies use active_cred, but could now offer the cached credential semantic used for the base system security model. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-19 19:04:53 +00:00
MAC_CHECK(check_vnode_read, active_cred, file_cred, vp,
&vp->v_label);
Break out mac_check_vnode_op() into three seperate checks: mac_check_vnode_poll(), mac_check_vnode_read(), mac_check_vnode_write(). This improves the consistency with other existing vnode checks, and allows policies to avoid implementing switch statements to determine what operations they do and do not want to authorize. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-19 16:43:25 +00:00
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
return (error);
}
int
mac_check_vnode_readdir(struct ucred *cred, struct vnode *dvp)
{
int error;
ASSERT_VOP_LOCKED(dvp, "mac_check_vnode_readdir");
if (!mac_enforce_fs)
return (0);
MAC_CHECK(check_vnode_readdir, cred, dvp, &dvp->v_label);
return (error);
}
int
mac_check_vnode_readlink(struct ucred *cred, struct vnode *vp)
{
int error;
ASSERT_VOP_LOCKED(vp, "mac_check_vnode_readlink");
if (!mac_enforce_fs)
return (0);
MAC_CHECK(check_vnode_readlink, cred, vp, &vp->v_label);
return (error);
}
static int
mac_check_vnode_relabel(struct ucred *cred, struct vnode *vp,
struct label *newlabel)
{
int error;
ASSERT_VOP_LOCKED(vp, "mac_check_vnode_relabel");
MAC_CHECK(check_vnode_relabel, cred, vp, &vp->v_label, newlabel);
return (error);
}
int
mac_check_vnode_rename_from(struct ucred *cred, struct vnode *dvp,
struct vnode *vp, struct componentname *cnp)
{
int error;
ASSERT_VOP_LOCKED(dvp, "mac_check_vnode_rename_from");
ASSERT_VOP_LOCKED(vp, "mac_check_vnode_rename_from");
if (!mac_enforce_fs)
return (0);
MAC_CHECK(check_vnode_rename_from, cred, dvp, &dvp->v_label, vp,
&vp->v_label, cnp);
return (error);
}
int
mac_check_vnode_rename_to(struct ucred *cred, struct vnode *dvp,
struct vnode *vp, int samedir, struct componentname *cnp)
{
int error;
ASSERT_VOP_LOCKED(dvp, "mac_check_vnode_rename_to");
ASSERT_VOP_LOCKED(vp, "mac_check_vnode_rename_to");
if (!mac_enforce_fs)
return (0);
MAC_CHECK(check_vnode_rename_to, cred, dvp, &dvp->v_label, vp,
vp != NULL ? &vp->v_label : NULL, samedir, cnp);
return (error);
}
int
mac_check_vnode_revoke(struct ucred *cred, struct vnode *vp)
{
int error;
ASSERT_VOP_LOCKED(vp, "mac_check_vnode_revoke");
if (!mac_enforce_fs)
return (0);
MAC_CHECK(check_vnode_revoke, cred, vp, &vp->v_label);
return (error);
}
int
mac_check_vnode_setacl(struct ucred *cred, struct vnode *vp, acl_type_t type,
struct acl *acl)
{
int error;
ASSERT_VOP_LOCKED(vp, "mac_check_vnode_setacl");
if (!mac_enforce_fs)
return (0);
MAC_CHECK(check_vnode_setacl, cred, vp, &vp->v_label, type, acl);
return (error);
}
int
mac_check_vnode_setextattr(struct ucred *cred, struct vnode *vp,
int attrnamespace, const char *name, struct uio *uio)
{
int error;
ASSERT_VOP_LOCKED(vp, "mac_check_vnode_setextattr");
if (!mac_enforce_fs)
return (0);
MAC_CHECK(check_vnode_setextattr, cred, vp, &vp->v_label,
attrnamespace, name, uio);
return (error);
}
int
mac_check_vnode_setflags(struct ucred *cred, struct vnode *vp, u_long flags)
{
int error;
ASSERT_VOP_LOCKED(vp, "mac_check_vnode_setflags");
if (!mac_enforce_fs)
return (0);
MAC_CHECK(check_vnode_setflags, cred, vp, &vp->v_label, flags);
return (error);
}
int
mac_check_vnode_setmode(struct ucred *cred, struct vnode *vp, mode_t mode)
{
int error;
ASSERT_VOP_LOCKED(vp, "mac_check_vnode_setmode");
if (!mac_enforce_fs)
return (0);
MAC_CHECK(check_vnode_setmode, cred, vp, &vp->v_label, mode);
return (error);
}
int
mac_check_vnode_setowner(struct ucred *cred, struct vnode *vp, uid_t uid,
gid_t gid)
{
int error;
ASSERT_VOP_LOCKED(vp, "mac_check_vnode_setowner");
if (!mac_enforce_fs)
return (0);
MAC_CHECK(check_vnode_setowner, cred, vp, &vp->v_label, uid, gid);
return (error);
}
int
mac_check_vnode_setutimes(struct ucred *cred, struct vnode *vp,
struct timespec atime, struct timespec mtime)
{
int error;
ASSERT_VOP_LOCKED(vp, "mac_check_vnode_setutimes");
if (!mac_enforce_fs)
return (0);
MAC_CHECK(check_vnode_setutimes, cred, vp, &vp->v_label, atime,
mtime);
return (error);
}
int
Pass active_cred and file_cred into the MAC framework explicitly for mac_check_vnode_{poll,read,stat,write}(). Pass in fp->f_cred when calling these checks with a struct file available. Otherwise, pass NOCRED. All currently MAC policies use active_cred, but could now offer the cached credential semantic used for the base system security model. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-19 19:04:53 +00:00
mac_check_vnode_stat(struct ucred *active_cred, struct ucred *file_cred,
struct vnode *vp)
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
{
int error;
ASSERT_VOP_LOCKED(vp, "mac_check_vnode_stat");
if (!mac_enforce_fs)
return (0);
Pass active_cred and file_cred into the MAC framework explicitly for mac_check_vnode_{poll,read,stat,write}(). Pass in fp->f_cred when calling these checks with a struct file available. Otherwise, pass NOCRED. All currently MAC policies use active_cred, but could now offer the cached credential semantic used for the base system security model. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-19 19:04:53 +00:00
MAC_CHECK(check_vnode_stat, active_cred, file_cred, vp,
&vp->v_label);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
return (error);
}
Break out mac_check_vnode_op() into three seperate checks: mac_check_vnode_poll(), mac_check_vnode_read(), mac_check_vnode_write(). This improves the consistency with other existing vnode checks, and allows policies to avoid implementing switch statements to determine what operations they do and do not want to authorize. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-19 16:43:25 +00:00
int
Pass active_cred and file_cred into the MAC framework explicitly for mac_check_vnode_{poll,read,stat,write}(). Pass in fp->f_cred when calling these checks with a struct file available. Otherwise, pass NOCRED. All currently MAC policies use active_cred, but could now offer the cached credential semantic used for the base system security model. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-19 19:04:53 +00:00
mac_check_vnode_write(struct ucred *active_cred, struct ucred *file_cred,
struct vnode *vp)
Break out mac_check_vnode_op() into three seperate checks: mac_check_vnode_poll(), mac_check_vnode_read(), mac_check_vnode_write(). This improves the consistency with other existing vnode checks, and allows policies to avoid implementing switch statements to determine what operations they do and do not want to authorize. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-19 16:43:25 +00:00
{
int error;
ASSERT_VOP_LOCKED(vp, "mac_check_vnode_write");
if (!mac_enforce_fs)
return (0);
Pass active_cred and file_cred into the MAC framework explicitly for mac_check_vnode_{poll,read,stat,write}(). Pass in fp->f_cred when calling these checks with a struct file available. Otherwise, pass NOCRED. All currently MAC policies use active_cred, but could now offer the cached credential semantic used for the base system security model. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-19 19:04:53 +00:00
MAC_CHECK(check_vnode_write, active_cred, file_cred, vp,
&vp->v_label);
Break out mac_check_vnode_op() into three seperate checks: mac_check_vnode_poll(), mac_check_vnode_read(), mac_check_vnode_write(). This improves the consistency with other existing vnode checks, and allows policies to avoid implementing switch statements to determine what operations they do and do not want to authorize. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-19 16:43:25 +00:00
return (error);
}
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
/*
* When relabeling a process, call out to the policies for the maximum
* permission allowed for each object type we know about in its
* memory space, and revoke access (in the least surprising ways we
* know) when necessary. The process lock is not held here.
*/
Un-staticize mac_cred_mmapped_drop_perms() so that it may be used by policy modules making use of downgrades in the MAC AST event. This is required by the mac_lomac port of LOMAC to the MAC Framework. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-11-26 17:11:57 +00:00
void
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
mac_cred_mmapped_drop_perms(struct thread *td, struct ucred *cred)
{
/* XXX freeze all other threads */
mac_cred_mmapped_drop_perms_recurse(td, cred,
&td->td_proc->p_vmspace->vm_map);
/* XXX allow other threads to continue */
}
static __inline const char *
prot2str(vm_prot_t prot)
{
switch (prot & VM_PROT_ALL) {
case VM_PROT_READ:
return ("r--");
case VM_PROT_READ | VM_PROT_WRITE:
return ("rw-");
case VM_PROT_READ | VM_PROT_EXECUTE:
return ("r-x");
case VM_PROT_READ | VM_PROT_WRITE | VM_PROT_EXECUTE:
return ("rwx");
case VM_PROT_WRITE:
return ("-w-");
case VM_PROT_EXECUTE:
return ("--x");
case VM_PROT_WRITE | VM_PROT_EXECUTE:
return ("-wx");
default:
return ("---");
}
}
static void
mac_cred_mmapped_drop_perms_recurse(struct thread *td, struct ucred *cred,
struct vm_map *map)
{
struct vm_map_entry *vme;
Sync from MAC tree: break out the single mmap entry point into seperate entry points for each occasion: mac_check_vnode_mmap() Check at initial mapping mac_check_vnode_mprotect() Check at mapping protection change mac_check_vnode_mmap_downgrade() Determine if a mapping downgrade should take place following subject relabel. Implement mmap() and mprotect() entry points for labeled vnode policies. These entry points are currently not hooked up to the VM system in the base tree. These changes improve the consistency of the access control interface and offer more flexibility regarding limiting access to vnode mmaping. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-06 02:46:26 +00:00
int result;
vm_prot_t revokeperms;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
vm_object_t object;
vm_ooffset_t offset;
struct vnode *vp;
Add security.mac.mmap_revocation, a flag indicating whether we should revoke access to memory maps on a process label change. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-09-09 17:12:24 +00:00
if (!mac_mmap_revocation)
return;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
vm_map_lock_read(map);
for (vme = map->header.next; vme != &map->header; vme = vme->next) {
if (vme->eflags & MAP_ENTRY_IS_SUB_MAP) {
mac_cred_mmapped_drop_perms_recurse(td, cred,
vme->object.sub_map);
continue;
}
/*
* Skip over entries that obviously are not shared.
*/
if (vme->eflags & (MAP_ENTRY_COW | MAP_ENTRY_NOSYNC) ||
!vme->max_protection)
continue;
/*
* Drill down to the deepest backing object.
*/
offset = vme->offset;
object = vme->object.vm_object;
if (object == NULL)
continue;
while (object->backing_object != NULL) {
object = object->backing_object;
offset += object->backing_object_offset;
}
/*
* At the moment, vm_maps and objects aren't considered
* by the MAC system, so only things with backing by a
* normal object (read: vnodes) are checked.
*/
if (object->type != OBJT_VNODE)
continue;
vp = (struct vnode *)object->handle;
vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, td);
Sync from MAC tree: break out the single mmap entry point into seperate entry points for each occasion: mac_check_vnode_mmap() Check at initial mapping mac_check_vnode_mprotect() Check at mapping protection change mac_check_vnode_mmap_downgrade() Determine if a mapping downgrade should take place following subject relabel. Implement mmap() and mprotect() entry points for labeled vnode policies. These entry points are currently not hooked up to the VM system in the base tree. These changes improve the consistency of the access control interface and offer more flexibility regarding limiting access to vnode mmaping. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-06 02:46:26 +00:00
result = vme->max_protection;
mac_check_vnode_mmap_downgrade(cred, vp, &result);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
VOP_UNLOCK(vp, 0, td);
/*
* Find out what maximum protection we may be allowing
* now but a policy needs to get removed.
*/
revokeperms = vme->max_protection & ~result;
if (!revokeperms)
continue;
Include <sys/malloc.h> instead of depending on namespace pollution 2 layers deep in <sys/proc.h> or <sys/vnode.h>. Removed unused includes. Fixed some printf format errors (1 fatal on i386's; 1 fatal on alphas; 1 not fatal on any supported machine).
2002-09-05 07:02:43 +00:00
printf("pid %ld: revoking %s perms from %#lx:%ld "
"(max %s/cur %s)\n", (long)td->td_proc->p_pid,
prot2str(revokeperms), (u_long)vme->start,
(long)(vme->end - vme->start),
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
prot2str(vme->max_protection), prot2str(vme->protection));
vm_map_lock_upgrade(map);
/*
* This is the really simple case: if a map has more
* max_protection than is allowed, but it's not being
* actually used (that is, the current protection is
* still allowed), we can just wipe it out and do
* nothing more.
*/
if ((vme->protection & revokeperms) == 0) {
vme->max_protection -= revokeperms;
} else {
if (revokeperms & VM_PROT_WRITE) {
/*
* In the more complicated case, flush out all
* pending changes to the object then turn it
* copy-on-write.
*/
vm_object_reference(object);
vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, td);
- Acquire the vm_object's lock when performing vm_object_page_clean(). - Add a parameter to vm_pageout_flush() that tells vm_pageout_flush() whether its caller has locked the vm_object. (This is a temporary measure to bootstrap vm_object locking.)
2003-04-24 04:31:25 +00:00
VM_OBJECT_LOCK(object);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
vm_object_page_clean(object,
OFF_TO_IDX(offset),
OFF_TO_IDX(offset + vme->end - vme->start +
PAGE_MASK),
OBJPC_SYNC);
- Acquire the vm_object's lock when performing vm_object_page_clean(). - Add a parameter to vm_pageout_flush() that tells vm_pageout_flush() whether its caller has locked the vm_object. (This is a temporary measure to bootstrap vm_object locking.)
2003-04-24 04:31:25 +00:00
VM_OBJECT_UNLOCK(object);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
VOP_UNLOCK(vp, 0, td);
vm_object_deallocate(object);
/*
* Why bother if there's no read permissions
* anymore? For the rest, we need to leave
* the write permissions on for COW, or
* remove them entirely if configured to.
*/
if (!mac_mmap_revocation_via_cow) {
vme->max_protection &= ~VM_PROT_WRITE;
vme->protection &= ~VM_PROT_WRITE;
} if ((revokeperms & VM_PROT_READ) == 0)
vme->eflags |= MAP_ENTRY_COW |
MAP_ENTRY_NEEDS_COPY;
}
if (revokeperms & VM_PROT_EXECUTE) {
vme->max_protection &= ~VM_PROT_EXECUTE;
vme->protection &= ~VM_PROT_EXECUTE;
}
if (revokeperms & VM_PROT_READ) {
vme->max_protection = 0;
vme->protection = 0;
}
pmap_protect(map->pmap, vme->start, vme->end,
vme->protection & ~revokeperms);
vm_map_simplify_entry(map, vme);
}
vm_map_lock_downgrade(map);
}
vm_map_unlock_read(map);
}
/*
* When the subject's label changes, it may require revocation of privilege
* to mapped objects. This can't be done on-the-fly later with a unified
* buffer cache.
*/
static void
mac_relabel_cred(struct ucred *cred, struct label *newlabel)
{
MAC_PERFORM(relabel_cred, cred, newlabel);
}
void
mac_relabel_vnode(struct ucred *cred, struct vnode *vp, struct label *newlabel)
{
MAC_PERFORM(relabel_vnode, cred, vp, &vp->v_label, newlabel);
}
void
mac_create_ifnet(struct ifnet *ifnet)
{
MAC_PERFORM(create_ifnet, ifnet, &ifnet->if_label);
}
void
mac_create_bpfdesc(struct ucred *cred, struct bpf_d *bpf_d)
{
MAC_PERFORM(create_bpfdesc, cred, bpf_d, &bpf_d->bd_label);
}
void
mac_create_socket(struct ucred *cred, struct socket *socket)
{
MAC_PERFORM(create_socket, cred, socket, &socket->so_label);
}
void
mac_create_pipe(struct ucred *cred, struct pipe *pipe)
{
MAC_PERFORM(create_pipe, cred, pipe, pipe->pipe_label);
}
void
mac_create_socket_from_socket(struct socket *oldsocket,
struct socket *newsocket)
{
MAC_PERFORM(create_socket_from_socket, oldsocket, &oldsocket->so_label,
newsocket, &newsocket->so_label);
}
static void
mac_relabel_socket(struct ucred *cred, struct socket *socket,
struct label *newlabel)
{
MAC_PERFORM(relabel_socket, cred, socket, &socket->so_label, newlabel);
}
static void
mac_relabel_pipe(struct ucred *cred, struct pipe *pipe, struct label *newlabel)
{
MAC_PERFORM(relabel_pipe, cred, pipe, pipe->pipe_label, newlabel);
}
void
mac_set_socket_peer_from_mbuf(struct mbuf *mbuf, struct socket *socket)
{
Abstract access to the mbuf header label behind a new function, mbuf_to_label(). This permits the vast majority of entry point code to be unaware that labels are stored in m->m_pkthdr.label, such that we can experiment storage of labels elsewhere (such as in m_tags). Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 18:11:18 +00:00
struct label *label;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
Abstract access to the mbuf header label behind a new function, mbuf_to_label(). This permits the vast majority of entry point code to be unaware that labels are stored in m->m_pkthdr.label, such that we can experiment storage of labels elsewhere (such as in m_tags). Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 18:11:18 +00:00
label = mbuf_to_label(mbuf);
MAC_PERFORM(set_socket_peer_from_mbuf, mbuf, label, socket,
&socket->so_peerlabel);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
void
mac_set_socket_peer_from_socket(struct socket *oldsocket,
struct socket *newsocket)
{
MAC_PERFORM(set_socket_peer_from_socket, oldsocket,
&oldsocket->so_label, newsocket, &newsocket->so_peerlabel);
}
void
mac_create_datagram_from_ipq(struct ipq *ipq, struct mbuf *datagram)
{
Abstract access to the mbuf header label behind a new function, mbuf_to_label(). This permits the vast majority of entry point code to be unaware that labels are stored in m->m_pkthdr.label, such that we can experiment storage of labels elsewhere (such as in m_tags). Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 18:11:18 +00:00
struct label *label;
label = mbuf_to_label(datagram);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
MAC_PERFORM(create_datagram_from_ipq, ipq, &ipq->ipq_label,
Abstract access to the mbuf header label behind a new function, mbuf_to_label(). This permits the vast majority of entry point code to be unaware that labels are stored in m->m_pkthdr.label, such that we can experiment storage of labels elsewhere (such as in m_tags). Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 18:11:18 +00:00
datagram, label);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
void
mac_create_fragment(struct mbuf *datagram, struct mbuf *fragment)
{
Abstract access to the mbuf header label behind a new function, mbuf_to_label(). This permits the vast majority of entry point code to be unaware that labels are stored in m->m_pkthdr.label, such that we can experiment storage of labels elsewhere (such as in m_tags). Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 18:11:18 +00:00
struct label *datagramlabel, *fragmentlabel;
datagramlabel = mbuf_to_label(datagram);
fragmentlabel = mbuf_to_label(fragment);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
Abstract access to the mbuf header label behind a new function, mbuf_to_label(). This permits the vast majority of entry point code to be unaware that labels are stored in m->m_pkthdr.label, such that we can experiment storage of labels elsewhere (such as in m_tags). Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 18:11:18 +00:00
MAC_PERFORM(create_fragment, datagram, datagramlabel, fragment,
fragmentlabel);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
void
mac_create_ipq(struct mbuf *fragment, struct ipq *ipq)
{
Abstract access to the mbuf header label behind a new function, mbuf_to_label(). This permits the vast majority of entry point code to be unaware that labels are stored in m->m_pkthdr.label, such that we can experiment storage of labels elsewhere (such as in m_tags). Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 18:11:18 +00:00
struct label *label;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
Abstract access to the mbuf header label behind a new function, mbuf_to_label(). This permits the vast majority of entry point code to be unaware that labels are stored in m->m_pkthdr.label, such that we can experiment storage of labels elsewhere (such as in m_tags). Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 18:11:18 +00:00
label = mbuf_to_label(fragment);
MAC_PERFORM(create_ipq, fragment, label, ipq, &ipq->ipq_label);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
void
mac_create_mbuf_from_mbuf(struct mbuf *oldmbuf, struct mbuf *newmbuf)
{
Abstract access to the mbuf header label behind a new function, mbuf_to_label(). This permits the vast majority of entry point code to be unaware that labels are stored in m->m_pkthdr.label, such that we can experiment storage of labels elsewhere (such as in m_tags). Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 18:11:18 +00:00
struct label *oldmbuflabel, *newmbuflabel;
oldmbuflabel = mbuf_to_label(oldmbuf);
newmbuflabel = mbuf_to_label(newmbuf);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
Abstract access to the mbuf header label behind a new function, mbuf_to_label(). This permits the vast majority of entry point code to be unaware that labels are stored in m->m_pkthdr.label, such that we can experiment storage of labels elsewhere (such as in m_tags). Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 18:11:18 +00:00
MAC_PERFORM(create_mbuf_from_mbuf, oldmbuf, oldmbuflabel, newmbuf,
newmbuflabel);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
void
mac_create_mbuf_from_bpfdesc(struct bpf_d *bpf_d, struct mbuf *mbuf)
{
Abstract access to the mbuf header label behind a new function, mbuf_to_label(). This permits the vast majority of entry point code to be unaware that labels are stored in m->m_pkthdr.label, such that we can experiment storage of labels elsewhere (such as in m_tags). Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 18:11:18 +00:00
struct label *label;
label = mbuf_to_label(mbuf);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
MAC_PERFORM(create_mbuf_from_bpfdesc, bpf_d, &bpf_d->bd_label, mbuf,
Abstract access to the mbuf header label behind a new function, mbuf_to_label(). This permits the vast majority of entry point code to be unaware that labels are stored in m->m_pkthdr.label, such that we can experiment storage of labels elsewhere (such as in m_tags). Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 18:11:18 +00:00
label);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
void
mac_create_mbuf_linklayer(struct ifnet *ifnet, struct mbuf *mbuf)
{
Abstract access to the mbuf header label behind a new function, mbuf_to_label(). This permits the vast majority of entry point code to be unaware that labels are stored in m->m_pkthdr.label, such that we can experiment storage of labels elsewhere (such as in m_tags). Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 18:11:18 +00:00
struct label *label;
label = mbuf_to_label(mbuf);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
MAC_PERFORM(create_mbuf_linklayer, ifnet, &ifnet->if_label, mbuf,
Abstract access to the mbuf header label behind a new function, mbuf_to_label(). This permits the vast majority of entry point code to be unaware that labels are stored in m->m_pkthdr.label, such that we can experiment storage of labels elsewhere (such as in m_tags). Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 18:11:18 +00:00
label);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
void
mac_create_mbuf_from_ifnet(struct ifnet *ifnet, struct mbuf *mbuf)
{
Abstract access to the mbuf header label behind a new function, mbuf_to_label(). This permits the vast majority of entry point code to be unaware that labels are stored in m->m_pkthdr.label, such that we can experiment storage of labels elsewhere (such as in m_tags). Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 18:11:18 +00:00
struct label *label;
label = mbuf_to_label(mbuf);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
MAC_PERFORM(create_mbuf_from_ifnet, ifnet, &ifnet->if_label, mbuf,
Abstract access to the mbuf header label behind a new function, mbuf_to_label(). This permits the vast majority of entry point code to be unaware that labels are stored in m->m_pkthdr.label, such that we can experiment storage of labels elsewhere (such as in m_tags). Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 18:11:18 +00:00
label);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
void
mac_create_mbuf_multicast_encap(struct mbuf *oldmbuf, struct ifnet *ifnet,
struct mbuf *newmbuf)
{
Abstract access to the mbuf header label behind a new function, mbuf_to_label(). This permits the vast majority of entry point code to be unaware that labels are stored in m->m_pkthdr.label, such that we can experiment storage of labels elsewhere (such as in m_tags). Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 18:11:18 +00:00
struct label *oldmbuflabel, *newmbuflabel;
oldmbuflabel = mbuf_to_label(oldmbuf);
newmbuflabel = mbuf_to_label(newmbuf);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
Abstract access to the mbuf header label behind a new function, mbuf_to_label(). This permits the vast majority of entry point code to be unaware that labels are stored in m->m_pkthdr.label, such that we can experiment storage of labels elsewhere (such as in m_tags). Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 18:11:18 +00:00
MAC_PERFORM(create_mbuf_multicast_encap, oldmbuf, oldmbuflabel,
ifnet, &ifnet->if_label, newmbuf, newmbuflabel);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
void
mac_create_mbuf_netlayer(struct mbuf *oldmbuf, struct mbuf *newmbuf)
{
Abstract access to the mbuf header label behind a new function, mbuf_to_label(). This permits the vast majority of entry point code to be unaware that labels are stored in m->m_pkthdr.label, such that we can experiment storage of labels elsewhere (such as in m_tags). Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 18:11:18 +00:00
struct label *oldmbuflabel, *newmbuflabel;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
Abstract access to the mbuf header label behind a new function, mbuf_to_label(). This permits the vast majority of entry point code to be unaware that labels are stored in m->m_pkthdr.label, such that we can experiment storage of labels elsewhere (such as in m_tags). Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 18:11:18 +00:00
oldmbuflabel = mbuf_to_label(oldmbuf);
newmbuflabel = mbuf_to_label(newmbuf);
MAC_PERFORM(create_mbuf_netlayer, oldmbuf, oldmbuflabel, newmbuf,
newmbuflabel);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
int
mac_fragment_match(struct mbuf *fragment, struct ipq *ipq)
{
Abstract access to the mbuf header label behind a new function, mbuf_to_label(). This permits the vast majority of entry point code to be unaware that labels are stored in m->m_pkthdr.label, such that we can experiment storage of labels elsewhere (such as in m_tags). Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 18:11:18 +00:00
struct label *label;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
int result;
Abstract access to the mbuf header label behind a new function, mbuf_to_label(). This permits the vast majority of entry point code to be unaware that labels are stored in m->m_pkthdr.label, such that we can experiment storage of labels elsewhere (such as in m_tags). Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 18:11:18 +00:00
label = mbuf_to_label(fragment);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
result = 1;
Abstract access to the mbuf header label behind a new function, mbuf_to_label(). This permits the vast majority of entry point code to be unaware that labels are stored in m->m_pkthdr.label, such that we can experiment storage of labels elsewhere (such as in m_tags). Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 18:11:18 +00:00
MAC_BOOLEAN(fragment_match, &&, fragment, label, ipq,
&ipq->ipq_label);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
return (result);
}
void
mac_update_ipq(struct mbuf *fragment, struct ipq *ipq)
{
Abstract access to the mbuf header label behind a new function, mbuf_to_label(). This permits the vast majority of entry point code to be unaware that labels are stored in m->m_pkthdr.label, such that we can experiment storage of labels elsewhere (such as in m_tags). Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 18:11:18 +00:00
struct label *label;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
Abstract access to the mbuf header label behind a new function, mbuf_to_label(). This permits the vast majority of entry point code to be unaware that labels are stored in m->m_pkthdr.label, such that we can experiment storage of labels elsewhere (such as in m_tags). Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 18:11:18 +00:00
label = mbuf_to_label(fragment);
MAC_PERFORM(update_ipq, fragment, label, ipq, &ipq->ipq_label);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
void
mac_create_mbuf_from_socket(struct socket *socket, struct mbuf *mbuf)
{
Abstract access to the mbuf header label behind a new function, mbuf_to_label(). This permits the vast majority of entry point code to be unaware that labels are stored in m->m_pkthdr.label, such that we can experiment storage of labels elsewhere (such as in m_tags). Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 18:11:18 +00:00
struct label *label;
label = mbuf_to_label(mbuf);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
MAC_PERFORM(create_mbuf_from_socket, socket, &socket->so_label, mbuf,
Abstract access to the mbuf header label behind a new function, mbuf_to_label(). This permits the vast majority of entry point code to be unaware that labels are stored in m->m_pkthdr.label, such that we can experiment storage of labels elsewhere (such as in m_tags). Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 18:11:18 +00:00
label);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
void
mac_create_mount(struct ucred *cred, struct mount *mp)
{
MAC_PERFORM(create_mount, cred, mp, &mp->mnt_mntlabel,
&mp->mnt_fslabel);
}
void
mac_create_root_mount(struct ucred *cred, struct mount *mp)
{
MAC_PERFORM(create_root_mount, cred, mp, &mp->mnt_mntlabel,
&mp->mnt_fslabel);
}
int
mac_check_bpfdesc_receive(struct bpf_d *bpf_d, struct ifnet *ifnet)
{
int error;
if (!mac_enforce_network)
return (0);
MAC_CHECK(check_bpfdesc_receive, bpf_d, &bpf_d->bd_label, ifnet,
&ifnet->if_label);
return (error);
}
static int
mac_check_cred_relabel(struct ucred *cred, struct label *newlabel)
{
int error;
MAC_CHECK(check_cred_relabel, cred, newlabel);
return (error);
}
int
mac_check_cred_visible(struct ucred *u1, struct ucred *u2)
{
int error;
if (!mac_enforce_process)
return (0);
MAC_CHECK(check_cred_visible, u1, u2);
return (error);
}
int
mac_check_ifnet_transmit(struct ifnet *ifnet, struct mbuf *mbuf)
{
Abstract access to the mbuf header label behind a new function, mbuf_to_label(). This permits the vast majority of entry point code to be unaware that labels are stored in m->m_pkthdr.label, such that we can experiment storage of labels elsewhere (such as in m_tags). Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 18:11:18 +00:00
struct label *label;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
int error;
Move MAC label storage for mbufs into m_tags from the m_pkthdr structure, returning some additional room in the first mbuf in a chain, and avoiding feature-specific contents in the mbuf header. To do this: - Modify mbuf_to_label() to extract the tag, returning NULL if not found. - Introduce mac_init_mbuf_tag() which does most of the work mac_init_mbuf() used to do, except on an m_tag rather than an mbuf. - Scale back mac_init_mbuf() to perform m_tag allocation and invoke mac_init_mbuf_tag(). - Replace mac_destroy_mbuf() with mac_destroy_mbuf_tag(), since m_tag's are now GC'd deep in the m_tag/mbuf code rather than at a higher level when mbufs are directly free()'d. - Add mac_copy_mbuf_tag() to support m_copy_pkthdr() and related notions. - Generally change all references to mbuf labels so that they use mbuf_to_label() rather than &mbuf->m_pkthdr.label. This required no changes in the MAC policies (yay!). - Tweak mbuf release routines to not call mac_destroy_mbuf(), tag destruction takes care of it for us now. - Remove MAC magic from m_copy_pkthdr() and m_move_pkthdr() -- the existing m_tag support does all this for us. Note that we can no longer just zero the m_tag list on the target mbuf, rather, we have to delete the chain because m_tag's will already be hung off freshly allocated mbuf's. - Tweak m_tag copying routines so that if we're copying a MAC m_tag, we don't do a binary copy, rather, we initialize the new storage and do a deep copy of the label. - Remove use of MAC_FLAG_INITIALIZED in a few bizarre places having to do with mbuf header copies previously. - When an mbuf is copied in ip_input(), we no longer need to explicitly copy the label because it will get handled by the m_tag code now. - No longer any weird handling of MAC labels in if_loop.c during header copies. - Add MPC_LOADTIME_FLAG_LABELMBUFS flag to Biba, MLS, mac_test. In mac_test, handle the label==NULL case, since it can be dynamically loaded. In order to improve performance with this change, introduce the notion of "lazy MAC label allocation" -- only allocate m_tag storage for MAC labels if we're running with a policy that uses MAC labels on mbufs. Policies declare this intent by setting the MPC_LOADTIME_FLAG_LABELMBUFS flag in their load-time flags field during declaration. Note: this opens up the possibility of post-boot policy modules getting back NULL slot entries even though they have policy invariants of non-NULL slot entries, as the policy might have been loaded after the mbuf was allocated, leaving the mbuf without label storage. Policies that cannot handle this case must be declared as NOTLATE, or must be modified. - mac_labelmbufs holds the current cumulative status as to whether any policies require mbuf labeling or not. This is updated whenever the active policy set changes by the function mac_policy_updateflags(). The function iterates the list and checks whether any have the flag set. Write access to this variable is protected by the policy list; read access is currently not protected for performance reasons. This might change if it causes problems. - Add MAC_POLICY_LIST_ASSERT_EXCLUSIVE() to permit the flags update function to assert appropriate locks. - This makes allocation in mac_init_mbuf() conditional on the flag. Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 20:39:06 +00:00
M_ASSERTPKTHDR(mbuf);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
if (!mac_enforce_network)
return (0);
Abstract access to the mbuf header label behind a new function, mbuf_to_label(). This permits the vast majority of entry point code to be unaware that labels are stored in m->m_pkthdr.label, such that we can experiment storage of labels elsewhere (such as in m_tags). Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 18:11:18 +00:00
label = mbuf_to_label(mbuf);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
MAC_CHECK(check_ifnet_transmit, ifnet, &ifnet->if_label, mbuf,
Abstract access to the mbuf header label behind a new function, mbuf_to_label(). This permits the vast majority of entry point code to be unaware that labels are stored in m->m_pkthdr.label, such that we can experiment storage of labels elsewhere (such as in m_tags). Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 18:11:18 +00:00
label);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
return (error);
}
Add MAC checks for various kenv() operations: dump, get, set, unset, permitting MAC policies to limit access to the kernel environment. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-11-01 20:46:53 +00:00
int
mac_check_kenv_dump(struct ucred *cred)
{
int error;
if (!mac_enforce_system)
return (0);
MAC_CHECK(check_kenv_dump, cred);
return (error);
}
int
mac_check_kenv_get(struct ucred *cred, char *name)
{
int error;
if (!mac_enforce_system)
return (0);
MAC_CHECK(check_kenv_get, cred, name);
return (error);
}
int
mac_check_kenv_set(struct ucred *cred, char *name, char *value)
{
int error;
if (!mac_enforce_system)
return (0);
MAC_CHECK(check_kenv_set, cred, name, value);
return (error);
}
int
mac_check_kenv_unset(struct ucred *cred, char *name)
{
int error;
if (!mac_enforce_system)
return (0);
MAC_CHECK(check_kenv_unset, cred, name);
return (error);
}
Merge kld access control checks from the MAC tree: these access control checks permit policy modules to augment the system policy for permitting kld operations. This permits policies to limit access to kld operations based on credential (and other) properties, as well as to perform checks on the kld being loaded (integrity, etc). Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-11-19 22:12:42 +00:00
int
mac_check_kld_load(struct ucred *cred, struct vnode *vp)
{
int error;
ASSERT_VOP_LOCKED(vp, "mac_check_kld_load");
if (!mac_enforce_kld)
return (0);
MAC_CHECK(check_kld_load, cred, vp, &vp->v_label);
return (error);
}
int
mac_check_kld_stat(struct ucred *cred)
{
int error;
if (!mac_enforce_kld)
return (0);
MAC_CHECK(check_kld_stat, cred);
return (error);
}
int
mac_check_kld_unload(struct ucred *cred)
{
int error;
if (!mac_enforce_kld)
return (0);
MAC_CHECK(check_kld_unload, cred);
return (error);
}
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
int
mac_check_mount_stat(struct ucred *cred, struct mount *mount)
{
int error;
if (!mac_enforce_fs)
return (0);
MAC_CHECK(check_mount_stat, cred, mount, &mount->mnt_mntlabel);
return (error);
}
int
mac_check_pipe_ioctl(struct ucred *cred, struct pipe *pipe, unsigned long cmd,
void *data)
{
int error;
Improve locking of pipe mutexes in the context of MAC: (1) Where previously the pipe mutex was selectively grabbed during pipe_ioctl(), now always grab it and then release if if not needed. This protects the call to mac_check_pipe_ioctl() to make sure the label remains consistent. (Note: it looks like sigio locking may be incorrect for fgetown() since we call it not-by-reference and sigio locking assumes call by reference). (2) In pipe_stat(), lock the pipe if MAC is compiled in so that the call to mac_check_pipe_stat() gets a locked pipe to protect label consistency. We still release the lock before returning actual stat() data, risking inconsistency, but apparently our pipe locking model accepts that risk. (3) In various pipe MAC authorization checks, assert that the pipe lock is held. (4) Grab the lock when performing a pipe relabel operation, and assert it a little deeper in the stack. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-01 04:30:19 +00:00
PIPE_LOCK_ASSERT(pipe, MA_OWNED);
if (!mac_enforce_pipe)
return (0);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
MAC_CHECK(check_pipe_ioctl, cred, pipe, pipe->pipe_label, cmd, data);
return (error);
}
int
Break out mac_check_pipe_op() into component check entry points: mac_check_pipe_poll(), mac_check_pipe_read(), mac_check_pipe_stat(), and mac_check_pipe_write(). This is improves consistency with other access control entry points and permits security modules to only control the object methods that they are interested in, avoiding switch statements. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-19 16:59:37 +00:00
mac_check_pipe_poll(struct ucred *cred, struct pipe *pipe)
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
{
int error;
Improve locking of pipe mutexes in the context of MAC: (1) Where previously the pipe mutex was selectively grabbed during pipe_ioctl(), now always grab it and then release if if not needed. This protects the call to mac_check_pipe_ioctl() to make sure the label remains consistent. (Note: it looks like sigio locking may be incorrect for fgetown() since we call it not-by-reference and sigio locking assumes call by reference). (2) In pipe_stat(), lock the pipe if MAC is compiled in so that the call to mac_check_pipe_stat() gets a locked pipe to protect label consistency. We still release the lock before returning actual stat() data, risking inconsistency, but apparently our pipe locking model accepts that risk. (3) In various pipe MAC authorization checks, assert that the pipe lock is held. (4) Grab the lock when performing a pipe relabel operation, and assert it a little deeper in the stack. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-01 04:30:19 +00:00
PIPE_LOCK_ASSERT(pipe, MA_OWNED);
if (!mac_enforce_pipe)
return (0);
Break out mac_check_pipe_op() into component check entry points: mac_check_pipe_poll(), mac_check_pipe_read(), mac_check_pipe_stat(), and mac_check_pipe_write(). This is improves consistency with other access control entry points and permits security modules to only control the object methods that they are interested in, avoiding switch statements. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-19 16:59:37 +00:00
MAC_CHECK(check_pipe_poll, cred, pipe, pipe->pipe_label);
return (error);
}
int
mac_check_pipe_read(struct ucred *cred, struct pipe *pipe)
{
int error;
Improve locking of pipe mutexes in the context of MAC: (1) Where previously the pipe mutex was selectively grabbed during pipe_ioctl(), now always grab it and then release if if not needed. This protects the call to mac_check_pipe_ioctl() to make sure the label remains consistent. (Note: it looks like sigio locking may be incorrect for fgetown() since we call it not-by-reference and sigio locking assumes call by reference). (2) In pipe_stat(), lock the pipe if MAC is compiled in so that the call to mac_check_pipe_stat() gets a locked pipe to protect label consistency. We still release the lock before returning actual stat() data, risking inconsistency, but apparently our pipe locking model accepts that risk. (3) In various pipe MAC authorization checks, assert that the pipe lock is held. (4) Grab the lock when performing a pipe relabel operation, and assert it a little deeper in the stack. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-01 04:30:19 +00:00
PIPE_LOCK_ASSERT(pipe, MA_OWNED);
if (!mac_enforce_pipe)
return (0);
Break out mac_check_pipe_op() into component check entry points: mac_check_pipe_poll(), mac_check_pipe_read(), mac_check_pipe_stat(), and mac_check_pipe_write(). This is improves consistency with other access control entry points and permits security modules to only control the object methods that they are interested in, avoiding switch statements. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-19 16:59:37 +00:00
MAC_CHECK(check_pipe_read, cred, pipe, pipe->pipe_label);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
return (error);
}
static int
mac_check_pipe_relabel(struct ucred *cred, struct pipe *pipe,
struct label *newlabel)
{
int error;
Improve locking of pipe mutexes in the context of MAC: (1) Where previously the pipe mutex was selectively grabbed during pipe_ioctl(), now always grab it and then release if if not needed. This protects the call to mac_check_pipe_ioctl() to make sure the label remains consistent. (Note: it looks like sigio locking may be incorrect for fgetown() since we call it not-by-reference and sigio locking assumes call by reference). (2) In pipe_stat(), lock the pipe if MAC is compiled in so that the call to mac_check_pipe_stat() gets a locked pipe to protect label consistency. We still release the lock before returning actual stat() data, risking inconsistency, but apparently our pipe locking model accepts that risk. (3) In various pipe MAC authorization checks, assert that the pipe lock is held. (4) Grab the lock when performing a pipe relabel operation, and assert it a little deeper in the stack. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-01 04:30:19 +00:00
PIPE_LOCK_ASSERT(pipe, MA_OWNED);
if (!mac_enforce_pipe)
return (0);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
MAC_CHECK(check_pipe_relabel, cred, pipe, pipe->pipe_label, newlabel);
return (error);
}
Break out mac_check_pipe_op() into component check entry points: mac_check_pipe_poll(), mac_check_pipe_read(), mac_check_pipe_stat(), and mac_check_pipe_write(). This is improves consistency with other access control entry points and permits security modules to only control the object methods that they are interested in, avoiding switch statements. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-19 16:59:37 +00:00
int
mac_check_pipe_stat(struct ucred *cred, struct pipe *pipe)
{
int error;
Improve locking of pipe mutexes in the context of MAC: (1) Where previously the pipe mutex was selectively grabbed during pipe_ioctl(), now always grab it and then release if if not needed. This protects the call to mac_check_pipe_ioctl() to make sure the label remains consistent. (Note: it looks like sigio locking may be incorrect for fgetown() since we call it not-by-reference and sigio locking assumes call by reference). (2) In pipe_stat(), lock the pipe if MAC is compiled in so that the call to mac_check_pipe_stat() gets a locked pipe to protect label consistency. We still release the lock before returning actual stat() data, risking inconsistency, but apparently our pipe locking model accepts that risk. (3) In various pipe MAC authorization checks, assert that the pipe lock is held. (4) Grab the lock when performing a pipe relabel operation, and assert it a little deeper in the stack. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-01 04:30:19 +00:00
PIPE_LOCK_ASSERT(pipe, MA_OWNED);
if (!mac_enforce_pipe)
return (0);
Break out mac_check_pipe_op() into component check entry points: mac_check_pipe_poll(), mac_check_pipe_read(), mac_check_pipe_stat(), and mac_check_pipe_write(). This is improves consistency with other access control entry points and permits security modules to only control the object methods that they are interested in, avoiding switch statements. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-19 16:59:37 +00:00
MAC_CHECK(check_pipe_stat, cred, pipe, pipe->pipe_label);
return (error);
}
int
mac_check_pipe_write(struct ucred *cred, struct pipe *pipe)
{
int error;
Improve locking of pipe mutexes in the context of MAC: (1) Where previously the pipe mutex was selectively grabbed during pipe_ioctl(), now always grab it and then release if if not needed. This protects the call to mac_check_pipe_ioctl() to make sure the label remains consistent. (Note: it looks like sigio locking may be incorrect for fgetown() since we call it not-by-reference and sigio locking assumes call by reference). (2) In pipe_stat(), lock the pipe if MAC is compiled in so that the call to mac_check_pipe_stat() gets a locked pipe to protect label consistency. We still release the lock before returning actual stat() data, risking inconsistency, but apparently our pipe locking model accepts that risk. (3) In various pipe MAC authorization checks, assert that the pipe lock is held. (4) Grab the lock when performing a pipe relabel operation, and assert it a little deeper in the stack. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-01 04:30:19 +00:00
PIPE_LOCK_ASSERT(pipe, MA_OWNED);
if (!mac_enforce_pipe)
return (0);
Break out mac_check_pipe_op() into component check entry points: mac_check_pipe_poll(), mac_check_pipe_read(), mac_check_pipe_stat(), and mac_check_pipe_write(). This is improves consistency with other access control entry points and permits security modules to only control the object methods that they are interested in, avoiding switch statements. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-19 16:59:37 +00:00
MAC_CHECK(check_pipe_write, cred, pipe, pipe->pipe_label);
return (error);
}
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
int
mac_check_proc_debug(struct ucred *cred, struct proc *proc)
{
int error;
Assert process locks in proces-related access control checks. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-19 15:30:30 +00:00
PROC_LOCK_ASSERT(proc, MA_OWNED);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
if (!mac_enforce_process)
return (0);
MAC_CHECK(check_proc_debug, cred, proc);
return (error);
}
int
mac_check_proc_sched(struct ucred *cred, struct proc *proc)
{
int error;
Assert process locks in proces-related access control checks. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-19 15:30:30 +00:00
PROC_LOCK_ASSERT(proc, MA_OWNED);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
if (!mac_enforce_process)
return (0);
MAC_CHECK(check_proc_sched, cred, proc);
return (error);
}
int
mac_check_proc_signal(struct ucred *cred, struct proc *proc, int signum)
{
int error;
Assert process locks in proces-related access control checks. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-19 15:30:30 +00:00
PROC_LOCK_ASSERT(proc, MA_OWNED);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
if (!mac_enforce_process)
return (0);
MAC_CHECK(check_proc_signal, cred, proc, signum);
return (error);
}
int
mac_check_socket_bind(struct ucred *ucred, struct socket *socket,
struct sockaddr *sockaddr)
{
int error;
if (!mac_enforce_socket)
return (0);
MAC_CHECK(check_socket_bind, ucred, socket, &socket->so_label,
sockaddr);
return (error);
}
int
mac_check_socket_connect(struct ucred *cred, struct socket *socket,
struct sockaddr *sockaddr)
{
int error;
if (!mac_enforce_socket)
return (0);
MAC_CHECK(check_socket_connect, cred, socket, &socket->so_label,
sockaddr);
return (error);
}
int
Rename mac_check_socket_receive() to mac_check_socket_deliver() so that we can use the names _receive() and _send() for the receive() and send() checks. Rename related constants, policy implementations, etc. PR: Submitted by: Reviewed by: Approved by: Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs MFC after:
2002-08-15 18:51:26 +00:00
mac_check_socket_deliver(struct socket *socket, struct mbuf *mbuf)
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
{
Abstract access to the mbuf header label behind a new function, mbuf_to_label(). This permits the vast majority of entry point code to be unaware that labels are stored in m->m_pkthdr.label, such that we can experiment storage of labels elsewhere (such as in m_tags). Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 18:11:18 +00:00
struct label *label;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
int error;
if (!mac_enforce_socket)
return (0);
Abstract access to the mbuf header label behind a new function, mbuf_to_label(). This permits the vast majority of entry point code to be unaware that labels are stored in m->m_pkthdr.label, such that we can experiment storage of labels elsewhere (such as in m_tags). Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 18:11:18 +00:00
label = mbuf_to_label(mbuf);
Rename mac_check_socket_receive() to mac_check_socket_deliver() so that we can use the names _receive() and _send() for the receive() and send() checks. Rename related constants, policy implementations, etc. PR: Submitted by: Reviewed by: Approved by: Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs MFC after:
2002-08-15 18:51:26 +00:00
MAC_CHECK(check_socket_deliver, socket, &socket->so_label, mbuf,
Abstract access to the mbuf header label behind a new function, mbuf_to_label(). This permits the vast majority of entry point code to be unaware that labels are stored in m->m_pkthdr.label, such that we can experiment storage of labels elsewhere (such as in m_tags). Reviewed by: sam Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-04-14 18:11:18 +00:00
label);
Rename mac_check_socket_receive() to mac_check_socket_deliver() so that we can use the names _receive() and _send() for the receive() and send() checks. Rename related constants, policy implementations, etc. PR: Submitted by: Reviewed by: Approved by: Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs MFC after:
2002-08-15 18:51:26 +00:00
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
return (error);
}
int
Rename mac_check_socket_receive() to mac_check_socket_deliver() so that we can use the names _receive() and _send() for the receive() and send() checks. Rename related constants, policy implementations, etc. PR: Submitted by: Reviewed by: Approved by: Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs MFC after:
2002-08-15 18:51:26 +00:00
mac_check_socket_listen(struct ucred *cred, struct socket *socket)
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
{
int error;
if (!mac_enforce_socket)
return (0);
Rename mac_check_socket_receive() to mac_check_socket_deliver() so that we can use the names _receive() and _send() for the receive() and send() checks. Rename related constants, policy implementations, etc. PR: Submitted by: Reviewed by: Approved by: Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs MFC after:
2002-08-15 18:51:26 +00:00
MAC_CHECK(check_socket_listen, cred, socket, &socket->so_label);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
return (error);
}
Integrate mac_check_socket_send() and mac_check_socket_receive() checks from the MAC tree: allow policies to perform access control for the ability of a process to send and receive data via a socket. At some point, we might also pass in additional address information if an explicit address is requested on send. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-06 14:39:15 +00:00
int
mac_check_socket_receive(struct ucred *cred, struct socket *so)
{
int error;
if (!mac_enforce_socket)
return (0);
MAC_CHECK(check_socket_receive, cred, so, &so->so_label);
return (error);
}
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
static int
mac_check_socket_relabel(struct ucred *cred, struct socket *socket,
struct label *newlabel)
{
int error;
MAC_CHECK(check_socket_relabel, cred, socket, &socket->so_label,
newlabel);
return (error);
}
Integrate mac_check_socket_send() and mac_check_socket_receive() checks from the MAC tree: allow policies to perform access control for the ability of a process to send and receive data via a socket. At some point, we might also pass in additional address information if an explicit address is requested on send. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-06 14:39:15 +00:00
int
mac_check_socket_send(struct ucred *cred, struct socket *so)
{
int error;
if (!mac_enforce_socket)
return (0);
MAC_CHECK(check_socket_send, cred, so, &so->so_label);
return (error);
}
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
int
mac_check_socket_visible(struct ucred *cred, struct socket *socket)
{
int error;
if (!mac_enforce_socket)
return (0);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
MAC_CHECK(check_socket_visible, cred, socket, &socket->so_label);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
return (error);
}
Instrument sysarch() MD privileged I/O access interfaces with a MAC check, mac_check_sysarch_ioperm(), permitting MAC security policy modules to control access to these interfaces. Currently, they protect access to IOPL on i386, and setting HAE on Alpha. Additional checks might be required on other platforms to prevent bypass of kernel security protections by unauthorized processes. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-03-06 04:47:47 +00:00
int
mac_check_sysarch_ioperm(struct ucred *cred)
{
int error;
if (!mac_enforce_system)
return (0);
MAC_CHECK(check_sysarch_ioperm, cred);
return (error);
}
Permit MAC policies to instrument the access control decisions for system accounting configuration and for nfsd server thread attach. Policies might use this to protect the integrity or confidentiality of accounting data, limit the ability to turn on or off accounting, as well as to prevent inappropriately labeled threads from becoming nfs server threads. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-11-04 15:13:36 +00:00
int
mac_check_system_acct(struct ucred *cred, struct vnode *vp)
{
int error;
if (vp != NULL) {
ASSERT_VOP_LOCKED(vp, "mac_check_system_acct");
}
if (!mac_enforce_system)
return (0);
MAC_CHECK(check_system_acct, cred, vp,
vp != NULL ? &vp->v_label : NULL);
return (error);
}
int
mac_check_system_nfsd(struct ucred *cred)
{
int error;
if (!mac_enforce_system)
return (0);
MAC_CHECK(check_system_nfsd, cred);
return (error);
}
Hook up mac_check_system_reboot(), a MAC Framework entry point that permits MAC modules to augment system security decisions regarding the reboot() system call, if MAC is compiled into the kernel. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-27 07:03:29 +00:00
int
mac_check_system_reboot(struct ucred *cred, int howto)
{
int error;
Centrally manage enforcement of {reboot,swapon,sysctl} using the mac_enforce_system toggle, rather than several separate toggles. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-27 15:50:49 +00:00
if (!mac_enforce_system)
Hook up mac_check_system_reboot(), a MAC Framework entry point that permits MAC modules to augment system security decisions regarding the reboot() system call, if MAC is compiled into the kernel. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-27 07:03:29 +00:00
return (0);
MAC_CHECK(check_system_reboot, cred, howto);
Centrally manage enforcement of {reboot,swapon,sysctl} using the mac_enforce_system toggle, rather than several separate toggles. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-27 15:50:49 +00:00
Hook up mac_check_system_reboot(), a MAC Framework entry point that permits MAC modules to augment system security decisions regarding the reboot() system call, if MAC is compiled into the kernel. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-27 07:03:29 +00:00
return (error);
}
Introduce mac_check_system_settime(), a MAC check allowing policies to augment the system policy for changing the system time. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-11-03 02:39:42 +00:00
int
mac_check_system_settime(struct ucred *cred)
{
int error;
if (!mac_enforce_system)
return (0);
MAC_CHECK(check_system_settime, cred);
return (error);
}
Merge from MAC tree: rename mac_check_vnode_swapon() to mac_check_system_swapon(), to reflect the fact that the primary object of this change is the running kernel as a whole, rather than just the vnode. We'll drop additional checks of this class into the same check namespace, including reboot(), sysctl(), et al. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-27 06:54:06 +00:00
int
mac_check_system_swapon(struct ucred *cred, struct vnode *vp)
{
int error;
ASSERT_VOP_LOCKED(vp, "mac_check_system_swapon");
Centrally manage enforcement of {reboot,swapon,sysctl} using the mac_enforce_system toggle, rather than several separate toggles. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-27 15:50:49 +00:00
if (!mac_enforce_system)
Merge from MAC tree: rename mac_check_vnode_swapon() to mac_check_system_swapon(), to reflect the fact that the primary object of this change is the running kernel as a whole, rather than just the vnode. We'll drop additional checks of this class into the same check namespace, including reboot(), sysctl(), et al. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-27 06:54:06 +00:00
return (0);
MAC_CHECK(check_system_swapon, cred, vp, &vp->v_label);
return (error);
}
Provide a mac_check_system_swapoff() entry point, which permits MAC modules to authorize disabling of swap against a particular vnode. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-03-05 23:50:15 +00:00
int
mac_check_system_swapoff(struct ucred *cred, struct vnode *vp)
{
int error;
ASSERT_VOP_LOCKED(vp, "mac_check_system_swapoff");
if (!mac_enforce_system)
return (0);
MAC_CHECK(check_system_swapoff, cred, vp, &vp->v_label);
return (error);
}
Implement mac_check_system_sysctl(), a MAC Framework entry point to permit MAC policies to augment the security protections on sysctl() operations. This is not really a wonderful entry point, as we only have access to the MIB of the target sysctl entry, rather than the more useful entry name, but this is sufficient for policies like Biba that wish to use their notions of privilege or integrity to prevent inappropriate sysctl modification. Affects MAC kernels only. Since SYSCTL_LOCK isn't in sysctl.h, just kern_sysctl.c, we can't assert the SYSCTL subsystem lockin the MAC Framework. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-27 07:12:34 +00:00
int
mac_check_system_sysctl(struct ucred *cred, int *name, u_int namelen,
void *old, size_t *oldlenp, int inkernel, void *new, size_t newlen)
{
int error;
/*
* XXXMAC: We're very much like to assert the SYSCTL_LOCK here,
* but since it's not exported from kern_sysctl.c, we can't.
*/
Centrally manage enforcement of {reboot,swapon,sysctl} using the mac_enforce_system toggle, rather than several separate toggles. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-27 15:50:49 +00:00
if (!mac_enforce_system)
Implement mac_check_system_sysctl(), a MAC Framework entry point to permit MAC policies to augment the security protections on sysctl() operations. This is not really a wonderful entry point, as we only have access to the MIB of the target sysctl entry, rather than the more useful entry name, but this is sufficient for policies like Biba that wish to use their notions of privilege or integrity to prevent inappropriate sysctl modification. Affects MAC kernels only. Since SYSCTL_LOCK isn't in sysctl.h, just kern_sysctl.c, we can't assert the SYSCTL subsystem lockin the MAC Framework. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-27 07:12:34 +00:00
return (0);
MAC_CHECK(check_system_sysctl, cred, name, namelen, old, oldlenp,
inkernel, new, newlen);
return (error);
}
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
int
mac_ioctl_ifnet_get(struct ucred *cred, struct ifreq *ifr,
struct ifnet *ifnet)
{
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
char *elements, *buffer;
struct mac mac;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
int error;
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
error = copyin(ifr->ifr_ifru.ifru_data, &mac, sizeof(mac));
if (error)
return (error);
error = mac_check_structmac_consistent(&mac);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
if (error)
return (error);
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
elements = malloc(mac.m_buflen, M_MACTEMP, M_WAITOK);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
error = copyinstr(mac.m_string, elements, mac.m_buflen, NULL);
if (error) {
free(elements, M_MACTEMP);
return (error);
}
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
buffer = malloc(mac.m_buflen, M_MACTEMP, M_WAITOK | M_ZERO);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
error = mac_externalize_ifnet_label(&ifnet->if_label, elements,
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
buffer, mac.m_buflen, M_WAITOK);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
if (error == 0)
error = copyout(buffer, mac.m_string, strlen(buffer)+1);
free(buffer, M_MACTEMP);
free(elements, M_MACTEMP);
return (error);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
int
mac_ioctl_ifnet_set(struct ucred *cred, struct ifreq *ifr,
struct ifnet *ifnet)
{
struct label intlabel;
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
struct mac mac;
char *buffer;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
int error;
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
error = copyin(ifr->ifr_ifru.ifru_data, &mac, sizeof(mac));
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
if (error)
return (error);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
error = mac_check_structmac_consistent(&mac);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
if (error)
return (error);
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
buffer = malloc(mac.m_buflen, M_MACTEMP, M_WAITOK);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
error = copyinstr(mac.m_string, buffer, mac.m_buflen, NULL);
if (error) {
free(buffer, M_MACTEMP);
return (error);
}
mac_init_ifnet_label(&intlabel);
error = mac_internalize_ifnet_label(&intlabel, buffer);
free(buffer, M_MACTEMP);
if (error) {
mac_destroy_ifnet_label(&intlabel);
return (error);
}
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
/*
* XXX: Note that this is a redundant privilege check, since
* policies impose this check themselves if required by the
* policy. Eventually, this should go away.
*/
error = suser_cred(cred, 0);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
if (error) {
mac_destroy_ifnet_label(&intlabel);
return (error);
}
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
MAC_CHECK(check_ifnet_relabel, cred, ifnet, &ifnet->if_label,
&intlabel);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
if (error) {
mac_destroy_ifnet_label(&intlabel);
return (error);
}
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
MAC_PERFORM(relabel_ifnet, cred, ifnet, &ifnet->if_label, &intlabel);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
mac_destroy_ifnet_label(&intlabel);
return (0);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
void
Remove dm_root entry from struct devfs_mount. It's never set, and is unused. Replace it with a dm_mount back-pointer to the struct mount that the devfs_mount is associated with. Export that pointer to MAC Framework entry points, where all current policies don't use the pointer. This permits the SEBSD port of SELinux's FLASK/TE to compile out-of-the-box on 5.0-CURRENT with full file system labeling support. Approved by: re (murray) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-12-09 03:44:28 +00:00
mac_create_devfs_device(struct mount *mp, dev_t dev, struct devfs_dirent *de)
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
{
Remove dm_root entry from struct devfs_mount. It's never set, and is unused. Replace it with a dm_mount back-pointer to the struct mount that the devfs_mount is associated with. Export that pointer to MAC Framework entry points, where all current policies don't use the pointer. This permits the SEBSD port of SELinux's FLASK/TE to compile out-of-the-box on 5.0-CURRENT with full file system labeling support. Approved by: re (murray) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-12-09 03:44:28 +00:00
MAC_PERFORM(create_devfs_device, mp, dev, de, &de->de_label);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
Integrate a devfs/MAC fix from the MAC tree: avoid a race condition during devfs VOP symlink creation by introducing a new entry point to determine the label of the devfs_dirent prior to allocation of a vnode for the symlink. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 18:40:10 +00:00
void
Remove dm_root entry from struct devfs_mount. It's never set, and is unused. Replace it with a dm_mount back-pointer to the struct mount that the devfs_mount is associated with. Export that pointer to MAC Framework entry points, where all current policies don't use the pointer. This permits the SEBSD port of SELinux's FLASK/TE to compile out-of-the-box on 5.0-CURRENT with full file system labeling support. Approved by: re (murray) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-12-09 03:44:28 +00:00
mac_create_devfs_symlink(struct ucred *cred, struct mount *mp,
struct devfs_dirent *dd, struct devfs_dirent *de)
Integrate a devfs/MAC fix from the MAC tree: avoid a race condition during devfs VOP symlink creation by introducing a new entry point to determine the label of the devfs_dirent prior to allocation of a vnode for the symlink. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 18:40:10 +00:00
{
Remove dm_root entry from struct devfs_mount. It's never set, and is unused. Replace it with a dm_mount back-pointer to the struct mount that the devfs_mount is associated with. Export that pointer to MAC Framework entry points, where all current policies don't use the pointer. This permits the SEBSD port of SELinux's FLASK/TE to compile out-of-the-box on 5.0-CURRENT with full file system labeling support. Approved by: re (murray) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-12-09 03:44:28 +00:00
MAC_PERFORM(create_devfs_symlink, cred, mp, dd, &dd->de_label, de,
Integrate a devfs/MAC fix from the MAC tree: avoid a race condition during devfs VOP symlink creation by introducing a new entry point to determine the label of the devfs_dirent prior to allocation of a vnode for the symlink. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-05 18:40:10 +00:00
&de->de_label);
}
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
void
Remove dm_root entry from struct devfs_mount. It's never set, and is unused. Replace it with a dm_mount back-pointer to the struct mount that the devfs_mount is associated with. Export that pointer to MAC Framework entry points, where all current policies don't use the pointer. This permits the SEBSD port of SELinux's FLASK/TE to compile out-of-the-box on 5.0-CURRENT with full file system labeling support. Approved by: re (murray) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-12-09 03:44:28 +00:00
mac_create_devfs_directory(struct mount *mp, char *dirname, int dirnamelen,
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
struct devfs_dirent *de)
{
Remove dm_root entry from struct devfs_mount. It's never set, and is unused. Replace it with a dm_mount back-pointer to the struct mount that the devfs_mount is associated with. Export that pointer to MAC Framework entry points, where all current policies don't use the pointer. This permits the SEBSD port of SELinux's FLASK/TE to compile out-of-the-box on 5.0-CURRENT with full file system labeling support. Approved by: re (murray) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-12-09 03:44:28 +00:00
MAC_PERFORM(create_devfs_directory, mp, dirname, dirnamelen, de,
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
&de->de_label);
}
int
mac_setsockopt_label_set(struct ucred *cred, struct socket *so,
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
struct mac *mac)
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
{
struct label intlabel;
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
char *buffer;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
int error;
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
error = mac_check_structmac_consistent(mac);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
if (error)
return (error);
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
buffer = malloc(mac->m_buflen, M_MACTEMP, M_WAITOK);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
error = copyinstr(mac->m_string, buffer, mac->m_buflen, NULL);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
if (error) {
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
free(buffer, M_MACTEMP);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
return (error);
}
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
mac_init_socket_label(&intlabel, M_WAITOK);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
error = mac_internalize_socket_label(&intlabel, buffer);
free(buffer, M_MACTEMP);
if (error) {
mac_destroy_socket_label(&intlabel);
return (error);
}
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
mac_check_socket_relabel(cred, so, &intlabel);
if (error) {
mac_destroy_socket_label(&intlabel);
return (error);
}
mac_relabel_socket(cred, so, &intlabel);
mac_destroy_socket_label(&intlabel);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
return (0);
}
int
mac_pipe_label_set(struct ucred *cred, struct pipe *pipe, struct label *label)
{
int error;
Improve locking of pipe mutexes in the context of MAC: (1) Where previously the pipe mutex was selectively grabbed during pipe_ioctl(), now always grab it and then release if if not needed. This protects the call to mac_check_pipe_ioctl() to make sure the label remains consistent. (Note: it looks like sigio locking may be incorrect for fgetown() since we call it not-by-reference and sigio locking assumes call by reference). (2) In pipe_stat(), lock the pipe if MAC is compiled in so that the call to mac_check_pipe_stat() gets a locked pipe to protect label consistency. We still release the lock before returning actual stat() data, risking inconsistency, but apparently our pipe locking model accepts that risk. (3) In various pipe MAC authorization checks, assert that the pipe lock is held. (4) Grab the lock when performing a pipe relabel operation, and assert it a little deeper in the stack. Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-01 04:30:19 +00:00
PIPE_LOCK_ASSERT(pipe, MA_OWNED);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
error = mac_check_pipe_relabel(cred, pipe, label);
if (error)
return (error);
mac_relabel_pipe(cred, pipe, label);
return (0);
}
int
mac_getsockopt_label_get(struct ucred *cred, struct socket *so,
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
struct mac *mac)
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
{
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
char *buffer, *elements;
int error;
error = mac_check_structmac_consistent(mac);
if (error)
return (error);
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
elements = malloc(mac->m_buflen, M_MACTEMP, M_WAITOK);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
error = copyinstr(mac->m_string, elements, mac->m_buflen, NULL);
if (error) {
free(elements, M_MACTEMP);
return (error);
}
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
buffer = malloc(mac->m_buflen, M_MACTEMP, M_WAITOK | M_ZERO);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
error = mac_externalize_socket_label(&so->so_label, elements,
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
buffer, mac->m_buflen, M_WAITOK);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
if (error == 0)
error = copyout(buffer, mac->m_string, strlen(buffer)+1);
free(buffer, M_MACTEMP);
free(elements, M_MACTEMP);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
return (error);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
int
mac_getsockopt_peerlabel_get(struct ucred *cred, struct socket *so,
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
struct mac *mac)
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
{
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
char *elements, *buffer;
int error;
error = mac_check_structmac_consistent(mac);
if (error)
return (error);
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
elements = malloc(mac->m_buflen, M_MACTEMP, M_WAITOK);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
error = copyinstr(mac->m_string, elements, mac->m_buflen, NULL);
if (error) {
free(elements, M_MACTEMP);
return (error);
}
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
buffer = malloc(mac->m_buflen, M_MACTEMP, M_WAITOK | M_ZERO);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
error = mac_externalize_socket_peer_label(&so->so_peerlabel,
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
elements, buffer, mac->m_buflen, M_WAITOK);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
if (error == 0)
error = copyout(buffer, mac->m_string, strlen(buffer)+1);
free(buffer, M_MACTEMP);
free(elements, M_MACTEMP);
return (error);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
/*
* Implementation of VOP_SETLABEL() that relies on extended attributes
* to store label data. Can be referenced by filesystems supporting
* extended attributes.
*/
int
vop_stdsetlabel_ea(struct vop_setlabel_args *ap)
{
struct vnode *vp = ap->a_vp;
struct label *intlabel = ap->a_label;
int error;
ASSERT_VOP_LOCKED(vp, "vop_stdsetlabel_ea");
Slightly change the semantics of vnode labels for MAC: rather than "refreshing" the label on the vnode before use, just get the label right from inception. For single-label file systems, set the label in the generic VFS getnewvnode() code; for multi-label file systems, leave the labeling up to the file system. With UFS1/2, this means reading the extended attribute during vfs_vget() as the inode is pulled off disk, rather than hitting the extended attributes frequently during operations later, improving performance. This also corrects sematics for shared vnode locks, which were not previously present in the system. This chances the cache coherrency properties WRT out-of-band access to label data, but in an acceptable form. With UFS1, there is a small race condition during automatic extended attribute start -- this is not present with UFS2, and occurs because EAs aren't available at vnode inception. We'll introduce a work around for this shortly. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-26 14:38:24 +00:00
if ((vp->v_mount->mnt_flag & MNT_MULTILABEL) == 0)
return (EOPNOTSUPP);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
Slightly change the semantics of vnode labels for MAC: rather than "refreshing" the label on the vnode before use, just get the label right from inception. For single-label file systems, set the label in the generic VFS getnewvnode() code; for multi-label file systems, leave the labeling up to the file system. With UFS1/2, this means reading the extended attribute during vfs_vget() as the inode is pulled off disk, rather than hitting the extended attributes frequently during operations later, improving performance. This also corrects sematics for shared vnode locks, which were not previously present in the system. This chances the cache coherrency properties WRT out-of-band access to label data, but in an acceptable form. With UFS1, there is a small race condition during automatic extended attribute start -- this is not present with UFS2, and occurs because EAs aren't available at vnode inception. We'll introduce a work around for this shortly. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-26 14:38:24 +00:00
error = mac_setlabel_vnode_extattr(ap->a_cred, vp, intlabel);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
if (error)
return (error);
mac_relabel_vnode(ap->a_cred, vp, intlabel);
return (0);
}
static int
vn_setlabel(struct vnode *vp, struct label *intlabel, struct ucred *cred)
{
int error;
if (vp->v_mount == NULL) {
/* printf("vn_setlabel: null v_mount\n"); */
Remove all use of vnode->v_tag, replacing with appropriate substitutes. v_tag is now const char * and should only be used for debugging. Additionally: 1. All users of VT_NTS now check vfsconf->vf_type VFCF_NETWORK 2. The user of VT_PROCFS now checks for the new flag VV_PROCDEP, which is propagated by pseudofs to all child vnodes if the fs sets PFS_PROCDEP. Suggested by: phk Reviewed by: bde, rwatson (earlier version)
2002-09-14 09:02:28 +00:00
if (vp->v_type != VNON)
printf("vn_setlabel: null v_mount with non-VNON\n");
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
return (EBADF);
}
if ((vp->v_mount->mnt_flag & MNT_MULTILABEL) == 0)
return (EOPNOTSUPP);
/*
* Multi-phase commit. First check the policies to confirm the
* change is OK. Then commit via the filesystem. Finally,
* update the actual vnode label. Question: maybe the filesystem
* should update the vnode at the end as part of VOP_SETLABEL()?
*/
error = mac_check_vnode_relabel(cred, vp, intlabel);
if (error)
return (error);
/*
* VADMIN provides the opportunity for the filesystem to make
* decisions about who is and is not able to modify labels
* and protections on files. This might not be right. We can't
* assume VOP_SETLABEL() will do it, because we might implement
* that as part of vop_stdsetlabel_ea().
*/
error = VOP_ACCESS(vp, VADMIN, cred, curthread);
if (error)
return (error);
error = VOP_SETLABEL(vp, intlabel, cred, curthread);
if (error)
return (error);
return (0);
}
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
int
__mac_get_pid(struct thread *td, struct __mac_get_pid_args *uap)
{
char *elements, *buffer;
struct mac mac;
struct proc *tproc;
struct ucred *tcred;
int error;
SCARGS removal take II.
2002-12-14 01:56:26 +00:00
error = copyin(uap->mac_p, &mac, sizeof(mac));
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
if (error)
return (error);
error = mac_check_structmac_consistent(&mac);
if (error)
return (error);
tproc = pfind(uap->pid);
if (tproc == NULL)
return (ESRCH);
tcred = NULL; /* Satisfy gcc. */
error = p_cansee(td, tproc);
if (error == 0)
tcred = crhold(tproc->p_ucred);
PROC_UNLOCK(tproc);
if (error)
return (error);
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
elements = malloc(mac.m_buflen, M_MACTEMP, M_WAITOK);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
error = copyinstr(mac.m_string, elements, mac.m_buflen, NULL);
if (error) {
free(elements, M_MACTEMP);
crfree(tcred);
return (error);
}
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
buffer = malloc(mac.m_buflen, M_MACTEMP, M_WAITOK | M_ZERO);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
error = mac_externalize_cred_label(&tcred->cr_label, elements,
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
buffer, mac.m_buflen, M_WAITOK);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
if (error == 0)
error = copyout(buffer, mac.m_string, strlen(buffer)+1);
free(buffer, M_MACTEMP);
free(elements, M_MACTEMP);
crfree(tcred);
return (error);
}
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
/*
* MPSAFE
*/
int
__mac_get_proc(struct thread *td, struct __mac_get_proc_args *uap)
{
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
char *elements, *buffer;
struct mac mac;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
int error;
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
error = copyin(uap->mac_p, &mac, sizeof(mac));
if (error)
return (error);
error = mac_check_structmac_consistent(&mac);
if (error)
return (error);
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
elements = malloc(mac.m_buflen, M_MACTEMP, M_WAITOK);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
error = copyinstr(mac.m_string, elements, mac.m_buflen, NULL);
if (error) {
free(elements, M_MACTEMP);
return (error);
}
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
buffer = malloc(mac.m_buflen, M_MACTEMP, M_WAITOK | M_ZERO);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
error = mac_externalize_cred_label(&td->td_ucred->cr_label,
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
elements, buffer, mac.m_buflen, M_WAITOK);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
if (error == 0)
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
error = copyout(buffer, mac.m_string, strlen(buffer)+1);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
free(buffer, M_MACTEMP);
free(elements, M_MACTEMP);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
return (error);
}
/*
* MPSAFE
*/
int
__mac_set_proc(struct thread *td, struct __mac_set_proc_args *uap)
{
struct ucred *newcred, *oldcred;
struct label intlabel;
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
struct proc *p;
struct mac mac;
char *buffer;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
int error;
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
error = copyin(uap->mac_p, &mac, sizeof(mac));
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
if (error)
return (error);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
error = mac_check_structmac_consistent(&mac);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
if (error)
return (error);
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
buffer = malloc(mac.m_buflen, M_MACTEMP, M_WAITOK);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
error = copyinstr(mac.m_string, buffer, mac.m_buflen, NULL);
if (error) {
free(buffer, M_MACTEMP);
return (error);
}
mac_init_cred_label(&intlabel);
error = mac_internalize_cred_label(&intlabel, buffer);
free(buffer, M_MACTEMP);
if (error) {
mac_destroy_cred_label(&intlabel);
return (error);
}
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
newcred = crget();
p = td->td_proc;
PROC_LOCK(p);
oldcred = p->p_ucred;
error = mac_check_cred_relabel(oldcred, &intlabel);
if (error) {
PROC_UNLOCK(p);
crfree(newcred);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
goto out;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
setsugid(p);
crcopy(newcred, oldcred);
mac_relabel_cred(newcred, &intlabel);
p->p_ucred = newcred;
Close a race in process label changing opened due to dropping the proc locking when revoking access to mmaps. Instead, perform this later once we've changed the process label (hold onto a reference to the new cred so that we don't lose it when we release the process lock if another thread changes the credential). Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-19 20:26:32 +00:00
/*
* Grab additional reference for use while revoking mmaps, prior
* to releasing the proc lock and sharing the cred.
*/
crhold(newcred);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
PROC_UNLOCK(p);
Close a race in process label changing opened due to dropping the proc locking when revoking access to mmaps. Instead, perform this later once we've changed the process label (hold onto a reference to the new cred so that we don't lose it when we release the process lock if another thread changes the credential). Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-19 20:26:32 +00:00
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
if (mac_enforce_vm) {
mtx_lock(&Giant);
mac_cred_mmapped_drop_perms(td, newcred);
mtx_unlock(&Giant);
}
Close a race in process label changing opened due to dropping the proc locking when revoking access to mmaps. Instead, perform this later once we've changed the process label (hold onto a reference to the new cred so that we don't lose it when we release the process lock if another thread changes the credential). Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-19 20:26:32 +00:00
crfree(newcred); /* Free revocation reference. */
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
crfree(oldcred);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
out:
mac_destroy_cred_label(&intlabel);
return (error);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
/*
* MPSAFE
*/
int
__mac_get_fd(struct thread *td, struct __mac_get_fd_args *uap)
{
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
char *elements, *buffer;
struct label intlabel;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
struct file *fp;
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
struct mac mac;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
struct vnode *vp;
struct pipe *pipe;
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
short label_type;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
int error;
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
error = copyin(uap->mac_p, &mac, sizeof(mac));
if (error)
return (error);
error = mac_check_structmac_consistent(&mac);
if (error)
return (error);
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
elements = malloc(mac.m_buflen, M_MACTEMP, M_WAITOK);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
error = copyinstr(mac.m_string, elements, mac.m_buflen, NULL);
if (error) {
free(elements, M_MACTEMP);
return (error);
}
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
buffer = malloc(mac.m_buflen, M_MACTEMP, M_WAITOK | M_ZERO);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
mtx_lock(&Giant); /* VFS */
SCARGS removal take II.
2002-12-14 01:56:26 +00:00
error = fget(td, uap->fd, &fp);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
if (error)
goto out;
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
label_type = fp->f_type;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
switch (fp->f_type) {
case DTYPE_FIFO:
case DTYPE_VNODE:
Add a f_vnode field to struct file. Several of the subtypes have an associated vnode which is used for stuff like the f*() functions. By giving the vnode a speparate field, a number of checks for the specific subtype can be replaced simply with a check for f_vnode != NULL, and we can later free f_data up to subtype specific use. At this point in time, f_data still points to the vnode, so any code I might have overlooked will still work.
2003-06-22 08:41:43 +00:00
vp = fp->f_vnode;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
mac_init_vnode_label(&intlabel);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, td);
Slightly change the semantics of vnode labels for MAC: rather than "refreshing" the label on the vnode before use, just get the label right from inception. For single-label file systems, set the label in the generic VFS getnewvnode() code; for multi-label file systems, leave the labeling up to the file system. With UFS1/2, this means reading the extended attribute during vfs_vget() as the inode is pulled off disk, rather than hitting the extended attributes frequently during operations later, improving performance. This also corrects sematics for shared vnode locks, which were not previously present in the system. This chances the cache coherrency properties WRT out-of-band access to label data, but in an acceptable form. With UFS1, there is a small race condition during automatic extended attribute start -- this is not present with UFS2, and occurs because EAs aren't available at vnode inception. We'll introduce a work around for this shortly. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-26 14:38:24 +00:00
mac_copy_vnode_label(&vp->v_label, &intlabel);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
VOP_UNLOCK(vp, 0, td);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
break;
case DTYPE_PIPE:
Bow to the whining masses and change a union back into void *. Retain removal of unnecessary casts and throw in some minor cleanups to see if anyone complains, just for the hell of it.
2003-01-13 00:33:17 +00:00
pipe = fp->f_data;
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
mac_init_pipe_label(&intlabel);
PIPE_LOCK(pipe);
mac_copy_pipe_label(pipe->pipe_label, &intlabel);
PIPE_UNLOCK(pipe);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
break;
default:
error = EINVAL;
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
fdrop(fp, td);
goto out;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
}
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
fdrop(fp, td);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
switch (label_type) {
case DTYPE_FIFO:
case DTYPE_VNODE:
if (error == 0)
error = mac_externalize_vnode_label(&intlabel,
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
elements, buffer, mac.m_buflen, M_WAITOK);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
mac_destroy_vnode_label(&intlabel);
break;
case DTYPE_PIPE:
error = mac_externalize_pipe_label(&intlabel, elements,
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
buffer, mac.m_buflen, M_WAITOK);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
mac_destroy_pipe_label(&intlabel);
break;
default:
panic("__mac_get_fd: corrupted label_type");
}
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
if (error == 0)
error = copyout(buffer, mac.m_string, strlen(buffer)+1);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
out:
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
mtx_unlock(&Giant); /* VFS */
free(buffer, M_MACTEMP);
free(elements, M_MACTEMP);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
return (error);
}
/*
* MPSAFE
*/
int
__mac_get_file(struct thread *td, struct __mac_get_file_args *uap)
{
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
char *elements, *buffer;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
struct nameidata nd;
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
struct label intlabel;
struct mac mac;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
int error;
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
error = copyin(uap->mac_p, &mac, sizeof(mac));
if (error)
return (error);
error = mac_check_structmac_consistent(&mac);
if (error)
return (error);
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
elements = malloc(mac.m_buflen, M_MACTEMP, M_WAITOK);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
error = copyinstr(mac.m_string, elements, mac.m_buflen, NULL);
if (error) {
free(elements, M_MACTEMP);
return (error);
}
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
buffer = malloc(mac.m_buflen, M_MACTEMP, M_WAITOK | M_ZERO);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
mtx_lock(&Giant); /* VFS */
NDINIT(&nd, LOOKUP, LOCKLEAF | FOLLOW, UIO_USERSPACE, uap->path_p,
td);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
error = namei(&nd);
if (error)
goto out;
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
mac_init_vnode_label(&intlabel);
Slightly change the semantics of vnode labels for MAC: rather than "refreshing" the label on the vnode before use, just get the label right from inception. For single-label file systems, set the label in the generic VFS getnewvnode() code; for multi-label file systems, leave the labeling up to the file system. With UFS1/2, this means reading the extended attribute during vfs_vget() as the inode is pulled off disk, rather than hitting the extended attributes frequently during operations later, improving performance. This also corrects sematics for shared vnode locks, which were not previously present in the system. This chances the cache coherrency properties WRT out-of-band access to label data, but in an acceptable form. With UFS1, there is a small race condition during automatic extended attribute start -- this is not present with UFS2, and occurs because EAs aren't available at vnode inception. We'll introduce a work around for this shortly. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-26 14:38:24 +00:00
mac_copy_vnode_label(&nd.ni_vp->v_label, &intlabel);
error = mac_externalize_vnode_label(&intlabel, elements, buffer,
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
mac.m_buflen, M_WAITOK);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
NDFREE(&nd, 0);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
mac_destroy_vnode_label(&intlabel);
if (error == 0)
error = copyout(buffer, mac.m_string, strlen(buffer)+1);
out:
mtx_unlock(&Giant); /* VFS */
free(buffer, M_MACTEMP);
free(elements, M_MACTEMP);
return (error);
}
/*
* MPSAFE
*/
int
__mac_get_link(struct thread *td, struct __mac_get_link_args *uap)
{
char *elements, *buffer;
struct nameidata nd;
struct label intlabel;
struct mac mac;
int error;
error = copyin(uap->mac_p, &mac, sizeof(mac));
if (error)
return (error);
error = mac_check_structmac_consistent(&mac);
if (error)
return (error);
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
elements = malloc(mac.m_buflen, M_MACTEMP, M_WAITOK);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
error = copyinstr(mac.m_string, elements, mac.m_buflen, NULL);
if (error) {
free(elements, M_MACTEMP);
return (error);
}
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
buffer = malloc(mac.m_buflen, M_MACTEMP, M_WAITOK | M_ZERO);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
mtx_lock(&Giant); /* VFS */
NDINIT(&nd, LOOKUP, LOCKLEAF | NOFOLLOW, UIO_USERSPACE, uap->path_p,
td);
error = namei(&nd);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
if (error)
goto out;
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
mac_init_vnode_label(&intlabel);
Slightly change the semantics of vnode labels for MAC: rather than "refreshing" the label on the vnode before use, just get the label right from inception. For single-label file systems, set the label in the generic VFS getnewvnode() code; for multi-label file systems, leave the labeling up to the file system. With UFS1/2, this means reading the extended attribute during vfs_vget() as the inode is pulled off disk, rather than hitting the extended attributes frequently during operations later, improving performance. This also corrects sematics for shared vnode locks, which were not previously present in the system. This chances the cache coherrency properties WRT out-of-band access to label data, but in an acceptable form. With UFS1, there is a small race condition during automatic extended attribute start -- this is not present with UFS2, and occurs because EAs aren't available at vnode inception. We'll introduce a work around for this shortly. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-26 14:38:24 +00:00
mac_copy_vnode_label(&nd.ni_vp->v_label, &intlabel);
error = mac_externalize_vnode_label(&intlabel, elements, buffer,
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
mac.m_buflen, M_WAITOK);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
NDFREE(&nd, 0);
mac_destroy_vnode_label(&intlabel);
if (error == 0)
error = copyout(buffer, mac.m_string, strlen(buffer)+1);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
out:
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
mtx_unlock(&Giant); /* VFS */
free(buffer, M_MACTEMP);
free(elements, M_MACTEMP);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
return (error);
}
/*
* MPSAFE
*/
int
__mac_set_fd(struct thread *td, struct __mac_set_fd_args *uap)
{
struct label intlabel;
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
struct pipe *pipe;
struct file *fp;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
struct mount *mp;
struct vnode *vp;
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
struct mac mac;
char *buffer;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
int error;
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
error = copyin(uap->mac_p, &mac, sizeof(mac));
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
if (error)
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
return (error);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
error = mac_check_structmac_consistent(&mac);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
if (error)
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
return (error);
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
buffer = malloc(mac.m_buflen, M_MACTEMP, M_WAITOK);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
error = copyinstr(mac.m_string, buffer, mac.m_buflen, NULL);
if (error) {
free(buffer, M_MACTEMP);
return (error);
}
mtx_lock(&Giant); /* VFS */
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
SCARGS removal take II.
2002-12-14 01:56:26 +00:00
error = fget(td, uap->fd, &fp);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
if (error)
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
goto out;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
switch (fp->f_type) {
case DTYPE_FIFO:
case DTYPE_VNODE:
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
mac_init_vnode_label(&intlabel);
error = mac_internalize_vnode_label(&intlabel, buffer);
if (error) {
mac_destroy_vnode_label(&intlabel);
break;
}
Add a f_vnode field to struct file. Several of the subtypes have an associated vnode which is used for stuff like the f*() functions. By giving the vnode a speparate field, a number of checks for the specific subtype can be replaced simply with a check for f_vnode != NULL, and we can later free f_data up to subtype specific use. At this point in time, f_data still points to the vnode, so any code I might have overlooked will still work.
2003-06-22 08:41:43 +00:00
vp = fp->f_vnode;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
error = vn_start_write(vp, &mp, V_WAIT | PCATCH);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
if (error != 0) {
mac_destroy_vnode_label(&intlabel);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
break;
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
}
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, td);
error = vn_setlabel(vp, &intlabel, td->td_ucred);
VOP_UNLOCK(vp, 0, td);
vn_finished_write(mp);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
mac_destroy_vnode_label(&intlabel);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
break;
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
case DTYPE_PIPE:
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
mac_init_pipe_label(&intlabel);
error = mac_internalize_pipe_label(&intlabel, buffer);
if (error == 0) {
Bow to the whining masses and change a union back into void *. Retain removal of unnecessary casts and throw in some minor cleanups to see if anyone complains, just for the hell of it.
2003-01-13 00:33:17 +00:00
pipe = fp->f_data;
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
PIPE_LOCK(pipe);
error = mac_pipe_label_set(td->td_ucred, pipe,
&intlabel);
PIPE_UNLOCK(pipe);
}
mac_destroy_pipe_label(&intlabel);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
break;
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
default:
error = EINVAL;
}
fdrop(fp, td);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
out:
mtx_unlock(&Giant); /* VFS */
free(buffer, M_MACTEMP);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
return (error);
}
/*
* MPSAFE
*/
int
__mac_set_file(struct thread *td, struct __mac_set_file_args *uap)
{
struct label intlabel;
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
struct nameidata nd;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
struct mount *mp;
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
struct mac mac;
char *buffer;
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
int error;
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
error = copyin(uap->mac_p, &mac, sizeof(mac));
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
if (error)
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
return (error);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
error = mac_check_structmac_consistent(&mac);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
if (error)
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
return (error);
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
buffer = malloc(mac.m_buflen, M_MACTEMP, M_WAITOK);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
error = copyinstr(mac.m_string, buffer, mac.m_buflen, NULL);
if (error) {
free(buffer, M_MACTEMP);
return (error);
}
mac_init_vnode_label(&intlabel);
error = mac_internalize_vnode_label(&intlabel, buffer);
free(buffer, M_MACTEMP);
if (error) {
mac_destroy_vnode_label(&intlabel);
return (error);
}
mtx_lock(&Giant); /* VFS */
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
NDINIT(&nd, LOOKUP, LOCKLEAF | FOLLOW, UIO_USERSPACE, uap->path_p,
td);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
error = namei(&nd);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
if (error == 0) {
error = vn_start_write(nd.ni_vp, &mp, V_WAIT | PCATCH);
if (error == 0)
error = vn_setlabel(nd.ni_vp, &intlabel,
td->td_ucred);
vn_finished_write(mp);
}
NDFREE(&nd, 0);
mtx_unlock(&Giant); /* VFS */
mac_destroy_vnode_label(&intlabel);
return (error);
}
/*
* MPSAFE
*/
int
__mac_set_link(struct thread *td, struct __mac_set_link_args *uap)
{
struct label intlabel;
struct nameidata nd;
struct mount *mp;
struct mac mac;
char *buffer;
int error;
error = copyin(uap->mac_p, &mac, sizeof(mac));
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
if (error)
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
return (error);
error = mac_check_structmac_consistent(&mac);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
if (error)
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
return (error);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
buffer = malloc(mac.m_buflen, M_MACTEMP, M_WAITOK);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
error = copyinstr(mac.m_string, buffer, mac.m_buflen, NULL);
if (error) {
free(buffer, M_MACTEMP);
return (error);
}
mac_init_vnode_label(&intlabel);
error = mac_internalize_vnode_label(&intlabel, buffer);
free(buffer, M_MACTEMP);
if (error) {
mac_destroy_vnode_label(&intlabel);
return (error);
}
mtx_lock(&Giant); /* VFS */
NDINIT(&nd, LOOKUP, LOCKLEAF | NOFOLLOW, UIO_USERSPACE, uap->path_p,
td);
error = namei(&nd);
if (error == 0) {
error = vn_start_write(nd.ni_vp, &mp, V_WAIT | PCATCH);
if (error == 0)
error = vn_setlabel(nd.ni_vp, &intlabel,
td->td_ucred);
vn_finished_write(mp);
}
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
NDFREE(&nd, 0);
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
mtx_unlock(&Giant); /* VFS */
mac_destroy_vnode_label(&intlabel);
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
return (error);
}
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
/*
* MPSAFE
*/
Provide an implementation of mac_syscall() so that security modules can offer new services without reserving system call numbers, or augmented versions of existing services. User code requests a target policy by name, and specifies the policy-specific API plus target. This is required in particular for our port of SELinux/FLASK to the MAC framework since it offers additional security services. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-19 17:59:48 +00:00
int
mac_syscall(struct thread *td, struct mac_syscall_args *uap)
{
struct mac_policy_conf *mpc;
char target[MAC_MAX_POLICY_NAME];
Clean up locking for the MAC Framework: (1) Accept that we're now going to use mutexes, so don't attempt to avoid treating them as mutexes. This cleans up locking accessor function names some. (2) Rename variables to _mtx, _cv, _count, simplifying the naming. (3) Add a new form of the _busy() primitive that conditionally makes the list busy: if there are entries on the list, bump the busy count. If there are no entries, don't bump the busy count. Return a boolean indicating whether or not the busy count was bumped. (4) Break mac_policy_list into two lists: one with the same name holding dynamic policies, and a new list, mac_static_policy_list, which holds policies loaded before mac_late and without the unload flag set. The static list may be accessed without holding the busy count, since it can't change at run-time. (5) In general, prefer making the list busy conditionally, meaning we pay only one mutex lock per entry point if all modules are on the static list, rather than two (since we don't have to lower the busy count when we're done with the framework). For systems running just Biba or MLS, this will halve the mutex accesses in the network stack, and may offer a substantial performance benefits. (6) Lay the groundwork for a dynamic-free kernel option which eliminates all locking associated with dynamically loaded or unloaded policies, for pre-configured systems requiring maximum performance but less run-time flexibility. These changes have been running for a few weeks on MAC development branch systems. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-07 17:49:24 +00:00
int entrycount, error;
Provide an implementation of mac_syscall() so that security modules can offer new services without reserving system call numbers, or augmented versions of existing services. User code requests a target policy by name, and specifies the policy-specific API plus target. This is required in particular for our port of SELinux/FLASK to the MAC framework since it offers additional security services. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-19 17:59:48 +00:00
SCARGS removal take II.
2002-12-14 01:56:26 +00:00
error = copyinstr(uap->policy, target, sizeof(target), NULL);
Provide an implementation of mac_syscall() so that security modules can offer new services without reserving system call numbers, or augmented versions of existing services. User code requests a target policy by name, and specifies the policy-specific API plus target. This is required in particular for our port of SELinux/FLASK to the MAC framework since it offers additional security services. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-19 17:59:48 +00:00
if (error)
return (error);
error = ENOSYS;
LIST_FOREACH(mpc, &mac_policy_list, mpc_list) {
if (strcmp(mpc->mpc_name, target) == 0 &&
mpc->mpc_ops->mpo_syscall != NULL) {
error = mpc->mpc_ops->mpo_syscall(td,
SCARGS removal take II.
2002-12-14 01:56:26 +00:00
uap->call, uap->arg);
Provide an implementation of mac_syscall() so that security modules can offer new services without reserving system call numbers, or augmented versions of existing services. User code requests a target policy by name, and specifies the policy-specific API plus target. This is required in particular for our port of SELinux/FLASK to the MAC framework since it offers additional security services. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-19 17:59:48 +00:00
goto out;
}
}
Clean up locking for the MAC Framework: (1) Accept that we're now going to use mutexes, so don't attempt to avoid treating them as mutexes. This cleans up locking accessor function names some. (2) Rename variables to _mtx, _cv, _count, simplifying the naming. (3) Add a new form of the _busy() primitive that conditionally makes the list busy: if there are entries on the list, bump the busy count. If there are no entries, don't bump the busy count. Return a boolean indicating whether or not the busy count was bumped. (4) Break mac_policy_list into two lists: one with the same name holding dynamic policies, and a new list, mac_static_policy_list, which holds policies loaded before mac_late and without the unload flag set. The static list may be accessed without holding the busy count, since it can't change at run-time. (5) In general, prefer making the list busy conditionally, meaning we pay only one mutex lock per entry point if all modules are on the static list, rather than two (since we don't have to lower the busy count when we're done with the framework). For systems running just Biba or MLS, this will halve the mutex accesses in the network stack, and may offer a substantial performance benefits. (6) Lay the groundwork for a dynamic-free kernel option which eliminates all locking associated with dynamically loaded or unloaded policies, for pre-configured systems requiring maximum performance but less run-time flexibility. These changes have been running for a few weeks on MAC development branch systems. Approved by: re (jhb) Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2003-05-07 17:49:24 +00:00
if ((entrycount = mac_policy_list_conditional_busy()) != 0) {
LIST_FOREACH(mpc, &mac_policy_list, mpc_list) {
if (strcmp(mpc->mpc_name, target) == 0 &&
mpc->mpc_ops->mpo_syscall != NULL) {
error = mpc->mpc_ops->mpo_syscall(td,
uap->call, uap->arg);
break;
}
}
mac_policy_list_unbusy();
}
Provide an implementation of mac_syscall() so that security modules can offer new services without reserving system call numbers, or augmented versions of existing services. User code requests a target policy by name, and specifies the policy-specific API plus target. This is required in particular for our port of SELinux/FLASK to the MAC framework since it offers additional security services. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-19 17:59:48 +00:00
out:
return (error);
}
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
SYSINIT(mac, SI_SUB_MAC, SI_ORDER_FIRST, mac_init, NULL);
SYSINIT(mac_late, SI_SUB_MAC_LATE, SI_ORDER_FIRST, mac_late_init, NULL);
#else /* !MAC */
Stubs for the TrustedBSD MAC system calls to permit TrustedBSD MAC userland code to operate on kernel's from the main tree. Not much in this file yet. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 02:04:05 +00:00
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
int
__mac_get_pid(struct thread *td, struct __mac_get_pid_args *uap)
{
return (ENOSYS);
}
Stubs for the TrustedBSD MAC system calls to permit TrustedBSD MAC userland code to operate on kernel's from the main tree. Not much in this file yet. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 02:04:05 +00:00
int
__mac_get_proc(struct thread *td, struct __mac_get_proc_args *uap)
{
return (ENOSYS);
}
int
__mac_set_proc(struct thread *td, struct __mac_set_proc_args *uap)
{
return (ENOSYS);
}
int
__mac_get_fd(struct thread *td, struct __mac_get_fd_args *uap)
{
return (ENOSYS);
}
int
__mac_get_file(struct thread *td, struct __mac_get_file_args *uap)
{
return (ENOSYS);
}
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
int
__mac_get_link(struct thread *td, struct __mac_get_link_args *uap)
{
return (ENOSYS);
}
Stubs for the TrustedBSD MAC system calls to permit TrustedBSD MAC userland code to operate on kernel's from the main tree. Not much in this file yet. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 02:04:05 +00:00
int
__mac_set_fd(struct thread *td, struct __mac_set_fd_args *uap)
{
return (ENOSYS);
}
int
__mac_set_file(struct thread *td, struct __mac_set_file_args *uap)
{
return (ENOSYS);
}
Begin committing support for Mandatory Access Control and extensible kernel access control. The MAC framework permits loadable kernel modules to link to the kernel at compile-time, boot-time, or run-time, and augment the system security policy. This commit includes the initial kernel implementation, although the interface with the userland components of the oeprating system is still under work, and not all kernel subsystems are supported. Later in this commit sequence, documentation of which kernel subsystems will not work correctly with a kernel compiled with MAC support will be added. kern_mac.c contains the body of the MAC framework. Kernel and user APIs defined in mac.h are implemented here, providing a front end to loaded security modules. This code implements a module registration service, state (label) management, security configuration and policy composition. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-07-30 21:36:05 +00:00
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
int
__mac_set_link(struct thread *td, struct __mac_set_link_args *uap)
{
return (ENOSYS);
}
Provide an implementation of mac_syscall() so that security modules can offer new services without reserving system call numbers, or augmented versions of existing services. User code requests a target policy by name, and specifies the policy-specific API plus target. This is required in particular for our port of SELinux/FLASK to the MAC framework since it offers additional security services. Obtained from: TrustedBSD Project Sponsored by: DARPA, NAI Labs
2002-08-19 17:59:48 +00:00
int
mac_syscall(struct thread *td, struct mac_syscall_args *uap)
{
return (ENOSYS);
}
Support the new MAC user API in kernel: modify existing system calls to use a modified notion of 'struct mac', and flesh out the new variation system calls (almost identical to existing ones except that they permit a pid to be specified for process label retrieval, and don't follow symlinks). This generalizes the label API so that the framework is now almost entirely policy-agnostic. Approved by: re Obtained from: TrustedBSD Project Sponsored by: DARPA, Network Associates Laboratories
2002-10-22 14:29:47 +00:00
#endif
Reference in New Issue Copy Permalink