| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
Add comprehensive KUnit tests for the namespace-related capability
functions that Serge Hallyn refactored in commit 9891d2f79a9f
("Clarify the rootid_owns_currentns").
The tests verify:
- Basic functionality: UID 0 in init namespace, invalid vfsuid,
non-zero UIDs
- Actual namespace traversal: Creating user namespaces with different
UID mappings where uid 0 maps to different kuids (e.g., 1000, 2000,
3000)
- Hierarchy traversal: Testing multiple nested namespaces to verify
correct namespace hierarchy traversal
This addresses the feedback to "test the actual functionality" by
creating real user namespaces with different values for the
namespace's uid 0, rather than just basic input validation.
The test file is included at the end of commoncap.c when
CONFIG_SECURITY_COMMONCAP_KUNIT_TEST is enabled, following the
standard kernel pattern (e.g., scsi_lib.c, ext4/mballoc.c). This
allows tests to access static functions in the same compilation unit
without modifying production code based on test configuration.
The tests require CONFIG_USER_NS to be enabled since they rely on user
namespace mapping functionality. The Kconfig dependency ensures the
tests only build when this requirement is met.
All 7 tests pass:
- test_vfsuid_root_in_currentns_init_ns
- test_vfsuid_root_in_currentns_invalid
- test_vfsuid_root_in_currentns_nonzero
- test_kuid_root_in_ns_init_ns_uid0
- test_kuid_root_in_ns_init_ns_nonzero
- test_kuid_root_in_ns_with_mapping
- test_kuid_root_in_ns_with_different_mappings
Updated MAINTAINER capabilities to include commoncap test
Signed-off-by: Ryan Foster <foster.ryan.r@gmail.com>
Signed-off-by: Serge Hallyn <sergeh@kernel.org>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/sergeh/linux
Pull capabilities update from Serge Hallyn:
"Ryan Foster had sent a patch to add testing of the
rootid_owns_currentns() function. That patch pointed out
that this function was not as clear as it should be. Fix it:
- Clarify the intent of the function in the name
- Split the function so that the base functionality is easier to test
from a kunit test"
* tag 'caps-pr-20251204' of git://git.kernel.org/pub/scm/linux/kernel/git/sergeh/linux:
Clarify the rootid_owns_currentns
|
|
Split most of the rootid_owns_currentns() functionality
into a more generic rootid_owns_ns() function which
will be easier to write tests for.
Rename the functions and variables to make clear that
the ids being tested could be any uid.
Signed-off-by: Serge Hallyn <serge@hallyn.com>
CC: Ryan Foster <foster.ryan.r@gmail.com>
CC: Christian Brauner <brauner@kernel.org>
---
v2: change the function parameter documentation to mollify the bot.
|
|
Reduce the duplication between the lsm_id struct and the DEFINE_LSM()
definition by linking the lsm_id struct directly into the individual
LSM's DEFINE_LSM() instance.
Linking the lsm_id into the LSM definition also allows us to simplify
the security_add_hooks() function by removing the code which populates
the lsm_idlist[] array and moving it into the normal LSM startup code
where the LSM list is parsed and the individual LSMs are enabled,
making for a cleaner implementation with less overhead at boot.
Reviewed-by: Kees Cook <kees@kernel.org>
Reviewed-by: John Johansen <john.johansen@canonical.com>
Reviewed-by: Casey Schaufler <casey@schaufler-ca.com>
Reviewed-by: Mimi Zohar <zohar@linux.ibm.com>
Signed-off-by: Paul Moore <paul@paul-moore.com>
|
|
Max Kellerman recently experienced a problem[1] when calling exec with
differing uid and euid's and he triggered the logic that is supposed
to only handle setuid executables.
When exec isn't changing anything in struct cred it doesn't make sense
to go into the code that is there to handle the case when the
credentials change.
When looking into the history of the code I discovered that this issue
was not present in Linux-2.4.0-test12 and was introduced in
Linux-2.4.0-prerelease when the logic for handling this case was moved
from prepare_binprm to compute_creds in fs/exec.c.
The bug introdused was to comparing euid in the new credentials with
uid instead of euid in the old credentials, when testing if setuid
had changed the euid.
Since triggering the keep ptrace limping along case for setuid
executables makes no sense when it was not a setuid exec revert back
to the logic present in Linux-2.4.0-test12.
This removes the confusingly named and subtlety incorrect helpers
is_setuid and is_setgid, that helped this bug to persist.
The varaiable is_setid is renamed to id_changed (it's Linux-2.4.0-test12)
as the old name describes what matters rather than it's cause.
The code removed in Linux-2.4.0-prerelease was:
- /* Set-uid? */
- if (mode & S_ISUID) {
- bprm->e_uid = inode->i_uid;
- if (bprm->e_uid != current->euid)
- id_change = 1;
- }
-
- /* Set-gid? */
- /*
- * If setgid is set but no group execute bit then this
- * is a candidate for mandatory locking, not a setgid
- * executable.
- */
- if ((mode & (S_ISGID | S_IXGRP)) == (S_ISGID | S_IXGRP)) {
- bprm->e_gid = inode->i_gid;
- if (!in_group_p(bprm->e_gid))
- id_change = 1;
Linux-2.4.0-prerelease added the current logic as:
+ if (bprm->e_uid != current->uid || bprm->e_gid != current->gid ||
+ !cap_issubset(new_permitted, current->cap_permitted)) {
+ current->dumpable = 0;
+
+ lock_kernel();
+ if (must_not_trace_exec(current)
+ || atomic_read(¤t->fs->count) > 1
+ || atomic_read(¤t->files->count) > 1
+ || atomic_read(¤t->sig->count) > 1) {
+ if(!capable(CAP_SETUID)) {
+ bprm->e_uid = current->uid;
+ bprm->e_gid = current->gid;
+ }
+ if(!capable(CAP_SETPCAP)) {
+ new_permitted = cap_intersect(new_permitted,
+ current->cap_permitted);
+ }
+ }
+ do_unlock = 1;
+ }
I have condenced the logic from Linux-2.4.0-test12 to just:
id_changed = !uid_eq(new->euid, old->euid) || !in_group_p(new->egid);
This change is userspace visible, but I don't expect anyone to care.
For the bug that is being fixed to trigger bprm->unsafe has to be set.
The variable bprm->unsafe is set when ptracing an executable, when
sharing a working directory, or when no_new_privs is set. Properly
testing for cases that are safe even in those conditions and doing
nothing special should not affect anyone. Especially if they were
previously ok with their credentials getting munged
To minimize behavioural changes the code continues to set secureexec
when euid != uid or when egid != gid.
[1] https://lkml.kernel.org/r/20250306082615.174777-1-max.kellermann@ionos.com
Reported-by: Max Kellermann <max.kellermann@ionos.com>
Fixes: 64444d3d0d7f ("Linux version 2.4.0-prerelease")
v1: https://lkml.kernel.org/r/878qmxsuy8.fsf@email.froward.int.ebiederm.org
Reviewed-by: Serge Hallyn <serge@hallyn.com>
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Reviewed-by: Jann Horn <jannh@google.com>
Acked-by: Kees Cook <kees@kernel.org>
|
|
The vanilla has_capability() function has been unused since 2018's
commit dcb569cf6ac9 ("Smack: ptrace capability use fixes")
Remove it.
Fixup a comment in security/commoncap.c that referenced it.
Signed-off-by: Dr. David Alan Gilbert <linux@treblig.org>
Reviewed-by: Paul Moore <paul@paul-moore.com>
Signed-off-by: Serge Hallyn <sergeh@kernel.org>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/sergeh/linux
Pull capabilities updates from Serge Hallyn:
- remove the cap_mmap_file() hook, as it simply returned the default
return value and so doesn't need to exist (Paul Moore)
- add a trace event for cap_capable() (Jordan Rome)
* tag 'caps-6.13-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/sergeh/linux:
security: add trace event for cap_capable
capabilities: remove cap_mmap_file()
|
|
The new SECBIT_EXEC_RESTRICT_FILE, SECBIT_EXEC_DENY_INTERACTIVE, and
their *_LOCKED counterparts are designed to be set by processes setting
up an execution environment, such as a user session, a container, or a
security sandbox. Unlike other securebits, these ones can be set by
unprivileged processes. Like seccomp filters or Landlock domains, the
securebits are inherited across processes.
When SECBIT_EXEC_RESTRICT_FILE is set, programs interpreting code should
control executable resources according to execveat(2) + AT_EXECVE_CHECK
(see previous commit).
When SECBIT_EXEC_DENY_INTERACTIVE is set, a process should deny
execution of user interactive commands (which excludes executable
regular files).
Being able to configure each of these securebits enables system
administrators or owner of image containers to gradually validate the
related changes and to identify potential issues (e.g. with interpreter
or audit logs).
It should be noted that unlike other security bits, the
SECBIT_EXEC_RESTRICT_FILE and SECBIT_EXEC_DENY_INTERACTIVE bits are
dedicated to user space willing to restrict itself. Because of that,
they only make sense in the context of a trusted environment (e.g.
sandbox, container, user session, full system) where the process
changing its behavior (according to these bits) and all its parent
processes are trusted. Otherwise, any parent process could just execute
its own malicious code (interpreting a script or not), or even enforce a
seccomp filter to mask these bits.
Such a secure environment can be achieved with an appropriate access
control (e.g. mount's noexec option, file access rights, LSM policy) and
an enlighten ld.so checking that libraries are allowed for execution
e.g., to protect against illegitimate use of LD_PRELOAD.
Ptrace restrictions according to these securebits would not make sense
because of the processes' trust assumption.
Scripts may need some changes to deal with untrusted data (e.g. stdin,
environment variables), but that is outside the scope of the kernel.
See chromeOS's documentation about script execution control and the
related threat model:
https://www.chromium.org/chromium-os/developer-library/guides/security/noexec-shell-scripts/
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Christian Brauner <brauner@kernel.org>
Cc: Kees Cook <keescook@chromium.org>
Cc: Paul Moore <paul@paul-moore.com>
Reviewed-by: Serge Hallyn <serge@hallyn.com>
Reviewed-by: Jeff Xu <jeffxu@chromium.org>
Tested-by: Jeff Xu <jeffxu@chromium.org>
Signed-off-by: Mickaël Salaün <mic@digikod.net>
Link: https://lore.kernel.org/r/20241212174223.389435-3-mic@digikod.net
Signed-off-by: Kees Cook <kees@kernel.org>
|
|
In cases where we want a stable way to observe/trace
cap_capable (e.g. protection from inlining and API updates)
add a tracepoint that passes:
- The credentials used
- The user namespace of the resource being accessed
- The user namespace in which the credential provides the
capability to access the targeted resource
- The capability to check for
- The return value of the check
Signed-off-by: Jordan Rome <linux@jordanrome.com>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Reviewed-by: Paul Moore <paul@paul-moore.com>
Reviewed-by: Serge Hallyn <serge@hallyn.com>
Link: https://lore.kernel.org/r/20241204155911.1817092-1-linux@jordanrome.com
Signed-off-by: Serge Hallyn <sergeh@kernel.org>
|
|
The cap_mmap_file() LSM callback returns the default value for the
security_mmap_file() LSM hook and can be safely removed.
Signed-off-by: Paul Moore <paul@paul-moore.com>
Reviewed-by: Casey Schaufler <casey@schaufler-ca.com>
Reviewed-by: Serge Hallyn <serge@hallyn.com>
Signed-off-by: Serge Hallyn <sergeh@kernel.org>
|
|
To be consistent with most LSM hooks, convert the return value of
hook vm_enough_memory to 0 or a negative error code.
Before:
- Hook vm_enough_memory returns 1 if permission is granted, 0 if not.
- LSM_RET_DEFAULT(vm_enough_memory_mm) is 1.
After:
- Hook vm_enough_memory reutrns 0 if permission is granted, negative
error code if not.
- LSM_RET_DEFAULT(vm_enough_memory_mm) is 0.
Signed-off-by: Xu Kuohai <xukuohai@huawei.com>
Reviewed-by: Casey Schaufler <casey@schaufler-ca.com>
Signed-off-by: Paul Moore <paul@paul-moore.com>
|
|
As the kernel test robot helpfully reminded us, all of the lsm_id
instances defined inside the various LSMs should be marked as static.
The one exception is Landlock which uses its lsm_id variable across
multiple source files with an extern declaration in a header file.
Reported-by: kernel test robot <lkp@intel.com>
Suggested-by: Casey Schaufler <casey@schaufler-ca.com>
Reviewed-by: Casey Schaufler <casey@schaufler-ca.com>
Signed-off-by: Paul Moore <paul@paul-moore.com>
|
|
Create a struct lsm_id to contain identifying information about Linux
Security Modules (LSMs). At inception this contains the name of the
module and an identifier associated with the security module. Change
the security_add_hooks() interface to use this structure. Change the
individual modules to maintain their own struct lsm_id and pass it to
security_add_hooks().
The values are for LSM identifiers are defined in a new UAPI
header file linux/lsm.h. Each existing LSM has been updated to
include it's LSMID in the lsm_id.
The LSM ID values are sequential, with the oldest module
LSM_ID_CAPABILITY being the lowest value and the existing modules
numbered in the order they were included in the main line kernel.
This is an arbitrary convention for assigning the values, but
none better presents itself. The value 0 is defined as being invalid.
The values 1-99 are reserved for any special case uses which may
arise in the future. This may include attributes of the LSM
infrastructure itself, possibly related to namespacing or network
attribute management. A special range is identified for such attributes
to help reduce confusion for developers unfamiliar with LSMs.
LSM attribute values are defined for the attributes presented by
modules that are available today. As with the LSM IDs, The value 0
is defined as being invalid. The values 1-99 are reserved for any
special case uses which may arise in the future.
Cc: linux-security-module <linux-security-module@vger.kernel.org>
Signed-off-by: Casey Schaufler <casey@schaufler-ca.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Serge Hallyn <serge@hallyn.com>
Reviewed-by: Mickael Salaun <mic@digikod.net>
Reviewed-by: John Johansen <john.johansen@canonical.com>
Signed-off-by: Kees Cook <keescook@chromium.org>
Nacked-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
[PM: forward ported beyond v6.6 due merge window changes]
Signed-off-by: Paul Moore <paul@paul-moore.com>
|
|
The 'bprm_creds_from_file' hook has implementation registered in
commoncap. Looking at the function implementation we observe that the
'file' parameter is not changing.
Mark the 'file' parameter of LSM hook security_bprm_creds_from_file() as
'const' since it will not be changing in the LSM hook.
Signed-off-by: Khadija Kamran <kamrankhadijadj@gmail.com>
Signed-off-by: Paul Moore <paul@paul-moore.com>
|
|
Three LSMs register the implementations for the "capget" hook: AppArmor,
SELinux, and the normal capability code. Looking at the function
implementations we may observe that the first parameter "target" is not
changing.
Mark the first argument "target" of LSM hook security_capget() as
"const" since it will not be changing in the LSM hook.
cap_capget() LSM hook declaration exceeds the 80 characters per line
limit. Split the function declaration to multiple lines to decrease the
line length.
Signed-off-by: Khadija Kamran <kamrankhadijadj@gmail.com>
Acked-by: John Johansen <john.johansen@canonical.com>
[PM: align the cap_capget() declaration, spelling fixes]
Signed-off-by: Paul Moore <paul@paul-moore.com>
|
|
A random collection of spelling fixes for source files in the LSM
layer.
Reviewed-by: Casey Schaufler <casey@schaufler-ca.com>
Signed-off-by: Paul Moore <paul@paul-moore.com>
|
|
After working with the larger SELinux-based distros for several
years, we're finally at a place where we can disable the SELinux
runtime disable functionality. The existing kernel deprecation
notice explains the functionality and why we want to remove it:
The selinuxfs "disable" node allows SELinux to be disabled at
runtime prior to a policy being loaded into the kernel. If
disabled via this mechanism, SELinux will remain disabled until
the system is rebooted.
The preferred method of disabling SELinux is via the "selinux=0"
boot parameter, but the selinuxfs "disable" node was created to
make it easier for systems with primitive bootloaders that did not
allow for easy modification of the kernel command line.
Unfortunately, allowing for SELinux to be disabled at runtime makes
it difficult to secure the kernel's LSM hooks using the
"__ro_after_init" feature.
It is that last sentence, mentioning the '__ro_after_init' hardening,
which is the real motivation for this change, and if you look at the
diffstat you'll see that the impact of this patch reaches across all
the different LSMs, helping prevent tampering at the LSM hook level.
From a SELinux perspective, it is important to note that if you
continue to disable SELinux via "/etc/selinux/config" it may appear
that SELinux is disabled, but it is simply in an uninitialized state.
If you load a policy with `load_policy -i`, you will see SELinux
come alive just as if you had loaded the policy during early-boot.
It is also worth noting that the "/sys/fs/selinux/disable" file is
always writable now, regardless of the Kconfig settings, but writing
to the file has no effect on the system, other than to display an
error on the console if a non-zero/true value is written.
Finally, in the several years where we have been working on
deprecating this functionality, there has only been one instance of
someone mentioning any user visible breakage. In this particular
case it was an individual's kernel test system, and the workaround
documented in the deprecation notice ("selinux=0" on the kernel
command line) resolved the issue without problem.
Acked-by: Casey Schaufler <casey@schaufler-ca.com>
Signed-off-by: Paul Moore <paul@paul-moore.com>
|
|
Back in 2008 we extended the capability bits from 32 to 64, and we did
it by extending the single 32-bit capability word from one word to an
array of two words. It was then obfuscated by hiding the "2" behind two
macro expansions, with the reasoning being that maybe it gets extended
further some day.
That reasoning may have been valid at the time, but the last thing we
want to do is to extend the capability set any more. And the array of
values not only causes source code oddities (with loops to deal with
it), but also results in worse code generation. It's a lose-lose
situation.
So just change the 'u32[2]' into a 'u64' and be done with it.
We still have to deal with the fact that the user space interface is
designed around an array of these 32-bit values, but that was the case
before too, since the array layouts were different (ie user space
doesn't use an array of 32-bit values for individual capability masks,
but an array of 32-bit slices of multiple masks).
So that marshalling of data is actually simplified too, even if it does
remain somewhat obscure and odd.
This was all triggered by my reaction to the new "cap_isidentical()"
introduced recently. By just using a saner data structure, it went from
unsigned __capi;
CAP_FOR_EACH_U32(__capi) {
if (a.cap[__capi] != b.cap[__capi])
return false;
}
return true;
to just being
return a.val == b.val;
instead. Which is rather more obvious both to humans and to compilers.
Cc: Mateusz Guzik <mjguzik@gmail.com>
Cc: Casey Schaufler <casey@schaufler-ca.com>
Cc: Serge Hallyn <serge@hallyn.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Paul Moore <paul@paul-moore.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
Convert to struct mnt_idmap.
Last cycle we merged the necessary infrastructure in
256c8aed2b42 ("fs: introduce dedicated idmap type for mounts").
This is just the conversion to struct mnt_idmap.
Currently we still pass around the plain namespace that was attached to a
mount. This is in general pretty convenient but it makes it easy to
conflate namespaces that are relevant on the filesystem with namespaces
that are relevent on the mount level. Especially for non-vfs developers
without detailed knowledge in this area this can be a potential source for
bugs.
Once the conversion to struct mnt_idmap is done all helpers down to the
really low-level helpers will take a struct mnt_idmap argument instead of
two namespace arguments. This way it becomes impossible to conflate the two
eliminating the possibility of any bugs. All of the vfs and all filesystems
only operate on struct mnt_idmap.
Acked-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner (Microsoft) <brauner@kernel.org>
|
|
Convert to struct mnt_idmap.
Last cycle we merged the necessary infrastructure in
256c8aed2b42 ("fs: introduce dedicated idmap type for mounts").
This is just the conversion to struct mnt_idmap.
Currently we still pass around the plain namespace that was attached to a
mount. This is in general pretty convenient but it makes it easy to
conflate namespaces that are relevant on the filesystem with namespaces
that are relevent on the mount level. Especially for non-vfs developers
without detailed knowledge in this area this can be a potential source for
bugs.
Once the conversion to struct mnt_idmap is done all helpers down to the
really low-level helpers will take a struct mnt_idmap argument instead of
two namespace arguments. This way it becomes impossible to conflate the two
eliminating the possibility of any bugs. All of the vfs and all filesystems
only operate on struct mnt_idmap.
Acked-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner (Microsoft) <brauner@kernel.org>
|
|
Convert to struct mnt_idmap.
Last cycle we merged the necessary infrastructure in
256c8aed2b42 ("fs: introduce dedicated idmap type for mounts").
This is just the conversion to struct mnt_idmap.
Currently we still pass around the plain namespace that was attached to a
mount. This is in general pretty convenient but it makes it easy to
conflate namespaces that are relevant on the filesystem with namespaces
that are relevent on the mount level. Especially for non-vfs developers
without detailed knowledge in this area this can be a potential source for
bugs.
Once the conversion to struct mnt_idmap is done all helpers down to the
really low-level helpers will take a struct mnt_idmap argument instead of
two namespace arguments. This way it becomes impossible to conflate the two
eliminating the possibility of any bugs. All of the vfs and all filesystems
only operate on struct mnt_idmap.
Acked-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner (Microsoft) <brauner@kernel.org>
|
|
Convert to struct mnt_idmap.
Last cycle we merged the necessary infrastructure in
256c8aed2b42 ("fs: introduce dedicated idmap type for mounts").
This is just the conversion to struct mnt_idmap.
Currently we still pass around the plain namespace that was attached to a
mount. This is in general pretty convenient but it makes it easy to
conflate namespaces that are relevant on the filesystem with namespaces
that are relevent on the mount level. Especially for non-vfs developers
without detailed knowledge in this area this can be a potential source for
bugs.
Once the conversion to struct mnt_idmap is done all helpers down to the
really low-level helpers will take a struct mnt_idmap argument instead of
two namespace arguments. This way it becomes impossible to conflate the two
eliminating the possibility of any bugs. All of the vfs and all filesystems
only operate on struct mnt_idmap.
Acked-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christian Brauner (Microsoft) <brauner@kernel.org>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/pcmoore/lsm
Pull lsm updates from Paul Moore:
- Improve the error handling in the device cgroup such that memory
allocation failures when updating the access policy do not
potentially alter the policy.
- Some minor fixes to reiserfs to ensure that it properly releases
LSM-related xattr values.
- Update the security_socket_getpeersec_stream() LSM hook to take
sockptr_t values.
Previously the net/BPF folks updated the getsockopt code in the
network stack to leverage the sockptr_t type to make it easier to
pass both kernel and __user pointers, but unfortunately when they did
so they didn't convert the LSM hook.
While there was/is no immediate risk by not converting the LSM hook,
it seems like this is a mistake waiting to happen so this patch
proactively does the LSM hook conversion.
- Convert vfs_getxattr_alloc() to return an int instead of a ssize_t
and cleanup the callers. Internally the function was never going to
return anything larger than an int and the callers were doing some
very odd things casting the return value; this patch fixes all that
and helps bring a bit of sanity to vfs_getxattr_alloc() and its
callers.
- More verbose, and helpful, LSM debug output when the system is booted
with "lsm.debug" on the command line. There are examples in the
commit description, but the quick summary is that this patch provides
better information about which LSMs are enabled and the ordering in
which they are processed.
- General comment and kernel-doc fixes and cleanups.
* tag 'lsm-pr-20221212' of git://git.kernel.org/pub/scm/linux/kernel/git/pcmoore/lsm:
lsm: Fix description of fs_context_parse_param
lsm: Add/fix return values in lsm_hooks.h and fix formatting
lsm: Clarify documentation of vm_enough_memory hook
reiserfs: Add missing calls to reiserfs_security_free()
lsm,fs: fix vfs_getxattr_alloc() return type and caller error paths
device_cgroup: Roll back to original exceptions after copy failure
LSM: Better reporting of actual LSMs at boot
lsm: make security_socket_getpeersec_stream() sockptr_t safe
audit: Fix some kernel-doc warnings
lsm: remove obsoleted comments for security hooks
fs: edit a comment made in bad taste
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/vfs/idmapping
Pull vfsuid updates from Christian Brauner:
"Last cycle we introduced the vfs{g,u}id_t types and associated helpers
to gain type safety when dealing with idmapped mounts. That initial
work already converted a lot of places over but there were still some
left,
This converts all remaining places that still make use of non-type
safe idmapping helpers to rely on the new type safe vfs{g,u}id based
helpers.
Afterwards it removes all the old non-type safe helpers"
* tag 'fs.vfsuid.conversion.v6.2' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/idmapping:
fs: remove unused idmapping helpers
ovl: port to vfs{g,u}id_t and associated helpers
fuse: port to vfs{g,u}id_t and associated helpers
ima: use type safe idmapping helpers
apparmor: use type safe idmapping helpers
caps: use type safe idmapping helpers
fs: use type safe idmapping helpers
mnt_idmapping: add missing helpers
|
|
The vfs_getxattr_alloc() function currently returns a ssize_t value
despite the fact that it only uses int values internally for return
values. Fix this by converting vfs_getxattr_alloc() to return an
int type and adjust the callers as necessary. As part of these
caller modifications, some of the callers are fixed to properly free
the xattr value buffer on both success and failure to ensure that
memory is not leaked in the failure case.
Reviewed-by: Serge Hallyn <serge@hallyn.com>
Reviewed-by: Mimi Zohar <zohar@linux.ibm.com>
Signed-off-by: Paul Moore <paul@paul-moore.com>
|
|
In cap_inode_getsecurity(), we will use vfs_getxattr_alloc() to
complete the memory allocation of tmpbuf, if we have completed
the memory allocation of tmpbuf, but failed to call handler->get(...),
there will be a memleak in below logic:
|-- ret = (int)vfs_getxattr_alloc(mnt_userns, ...)
| /* ^^^ alloc for tmpbuf */
|-- value = krealloc(*xattr_value, error + 1, flags)
| /* ^^^ alloc memory */
|-- error = handler->get(handler, ...)
| /* error! */
|-- *xattr_value = value
| /* xattr_value is &tmpbuf (memory leak!) */
So we will try to free(tmpbuf) after vfs_getxattr_alloc() fails to fix it.
Cc: stable@vger.kernel.org
Fixes: 8db6c34f1dbc ("Introduce v3 namespaced file capabilities")
Signed-off-by: Gaosheng Cui <cuigaosheng1@huawei.com>
Acked-by: Serge Hallyn <serge@hallyn.com>
[PM: subject line and backtrace tweaks]
Signed-off-by: Paul Moore <paul@paul-moore.com>
|
|
We already ported most parts and filesystems over for v6.0 to the new
vfs{g,u}id_t type and associated helpers for v6.0. Convert the remaining
places so we can remove all the old helpers.
This is a non-functional change.
Reviewed-by: Seth Forshee (DigitalOcean) <sforshee@kernel.org>
Signed-off-by: Christian Brauner (Microsoft) <brauner@kernel.org>
|
|
In previous patches we added new and modified existing helpers to handle
idmapped mounts of filesystems mounted with an idmapping. In this final
patch we convert all relevant places in the vfs to actually pass the
filesystem's idmapping into these helpers.
With this the vfs is in shape to handle idmapped mounts of filesystems
mounted with an idmapping. Note that this is just the generic
infrastructure. Actually adding support for idmapped mounts to a
filesystem mountable with an idmapping is follow-up work.
In this patch we extend the definition of an idmapped mount from a mount
that that has the initial idmapping attached to it to a mount that has
an idmapping attached to it which is not the same as the idmapping the
filesystem was mounted with.
As before we do not allow the initial idmapping to be attached to a
mount. In addition this patch prevents that the idmapping the filesystem
was mounted with can be attached to a mount created based on this
filesystem.
This has multiple reasons and advantages. First, attaching the initial
idmapping or the filesystem's idmapping doesn't make much sense as in
both cases the values of the i_{g,u}id and other places where k{g,u}ids
are used do not change. Second, a user that really wants to do this for
whatever reason can just create a separate dedicated identical idmapping
to attach to the mount. Third, we can continue to use the initial
idmapping as an indicator that a mount is not idmapped allowing us to
continue to keep passing the initial idmapping into the mapping helpers
to tell them that something isn't an idmapped mount even if the
filesystem is mounted with an idmapping.
Link: https://lore.kernel.org/r/20211123114227.3124056-11-brauner@kernel.org (v1)
Link: https://lore.kernel.org/r/20211130121032.3753852-11-brauner@kernel.org (v2)
Link: https://lore.kernel.org/r/20211203111707.3901969-11-brauner@kernel.org
Cc: Seth Forshee <sforshee@digitalocean.com>
Cc: Amir Goldstein <amir73il@gmail.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Al Viro <viro@zeniv.linux.org.uk>
CC: linux-fsdevel@vger.kernel.org
Reviewed-by: Seth Forshee <sforshee@digitalocean.com>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
|
|
In a few places the vfs needs to interact with bare k{g,u}ids directly
instead of struct inode. These are just a few. In previous patches we
introduced low-level mapping helpers that are able to support
filesystems mounted an idmapping. This patch simply converts the places
to use these new helpers.
Link: https://lore.kernel.org/r/20211123114227.3124056-7-brauner@kernel.org (v1)
Link: https://lore.kernel.org/r/20211130121032.3753852-7-brauner@kernel.org (v2)
Link: https://lore.kernel.org/r/20211203111707.3901969-7-brauner@kernel.org
Cc: Seth Forshee <sforshee@digitalocean.com>
Cc: Amir Goldstein <amir73il@gmail.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Al Viro <viro@zeniv.linux.org.uk>
CC: linux-fsdevel@vger.kernel.org
Reviewed-by: Seth Forshee <sforshee@digitalocean.com>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
|
|
The low-level mapping helpers were so far crammed into fs.h. They are
out of place there. The fs.h header should just contain the higher-level
mapping helpers that interact directly with vfs objects such as struct
super_block or struct inode and not the bare mapping helpers. Similarly,
only vfs and specific fs code shall interact with low-level mapping
helpers. And so they won't be made accessible automatically through
regular {g,u}id helpers.
Link: https://lore.kernel.org/r/20211123114227.3124056-3-brauner@kernel.org (v1)
Link: https://lore.kernel.org/r/20211130121032.3753852-3-brauner@kernel.org (v2)
Link: https://lore.kernel.org/r/20211203111707.3901969-3-brauner@kernel.org
Cc: Seth Forshee <sforshee@digitalocean.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Al Viro <viro@zeniv.linux.org.uk>
CC: linux-fsdevel@vger.kernel.org
Reviewed-by: Amir Goldstein <amir73il@gmail.com>
Reviewed-by: Seth Forshee <sforshee@digitalocean.com>
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/linux-security
Pull security layer fixes from James Morris:
"Miscellaneous minor fixes"
* tag 'fixes-v5.13' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/linux-security:
security: commoncap: clean up kernel-doc comments
security: commoncap: fix -Wstringop-overread warning
|
|
Fix kernel-doc notation in commoncap.c.
Use correct (matching) function name in comments as in code.
Use correct function argument names in kernel-doc comments.
Use kernel-doc's "Return:" format for function return values.
Fixes these kernel-doc warnings:
../security/commoncap.c:1206: warning: expecting prototype for cap_task_ioprio(). Prototype was for cap_task_setioprio() instead
../security/commoncap.c:1219: warning: expecting prototype for cap_task_ioprio(). Prototype was for cap_task_setnice() instead
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Reviewed-by: Serge Hallyn <serge@hallyn.com>
Signed-off-by: James Morris <jamorris@linux.microsoft.com>
|
|
gcc-11 introdces a harmless warning for cap_inode_getsecurity:
security/commoncap.c: In function ‘cap_inode_getsecurity’:
security/commoncap.c:440:33: error: ‘memcpy’ reading 16 bytes from a region of size 0 [-Werror=stringop-overread]
440 | memcpy(&nscap->data, &cap->data, sizeof(__le32) * 2 * VFS_CAP_U32);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The problem here is that tmpbuf is initialized to NULL, so gcc assumes
it is not accessible unless it gets set by vfs_getxattr_alloc(). This is
a legitimate warning as far as I can tell, but the code is correct since
it correctly handles the error when that function fails.
Add a separate NULL check to tell gcc about it as well.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Christian Brauner <christian.brauner@ubuntu.com>
Signed-off-by: James Morris <jamorris@linux.microsoft.com>
|
|
capabilities")
It turns out that there are in fact userspace implementations that
care and this recent change caused a regression.
https://github.com/containers/buildah/issues/3071
As the motivation for the original change was future development,
and the impact is existing real world code just revert this change
and allow the ambiguity in v3 file caps.
Cc: stable@vger.kernel.org
Fixes: 95ebabde382c ("capabilities: Don't allow writing ambiguous v3 file capabilities")
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/brauner/linux
Pull idmapped mounts from Christian Brauner:
"This introduces idmapped mounts which has been in the making for some
time. Simply put, different mounts can expose the same file or
directory with different ownership. This initial implementation comes
with ports for fat, ext4 and with Christoph's port for xfs with more
filesystems being actively worked on by independent people and
maintainers.
Idmapping mounts handle a wide range of long standing use-cases. Here
are just a few:
- Idmapped mounts make it possible to easily share files between
multiple users or multiple machines especially in complex
scenarios. For example, idmapped mounts will be used in the
implementation of portable home directories in
systemd-homed.service(8) where they allow users to move their home
directory to an external storage device and use it on multiple
computers where they are assigned different uids and gids. This
effectively makes it possible to assign random uids and gids at
login time.
- It is possible to share files from the host with unprivileged
containers without having to change ownership permanently through
chown(2).
- It is possible to idmap a container's rootfs and without having to
mangle every file. For example, Chromebooks use it to share the
user's Download folder with their unprivileged containers in their
Linux subsystem.
- It is possible to share files between containers with
non-overlapping idmappings.
- Filesystem that lack a proper concept of ownership such as fat can
use idmapped mounts to implement discretionary access (DAC)
permission checking.
- They allow users to efficiently changing ownership on a per-mount
basis without having to (recursively) chown(2) all files. In
contrast to chown (2) changing ownership of large sets of files is
instantenous with idmapped mounts. This is especially useful when
ownership of a whole root filesystem of a virtual machine or
container is changed. With idmapped mounts a single syscall
mount_setattr syscall will be sufficient to change the ownership of
all files.
- Idmapped mounts always take the current ownership into account as
idmappings specify what a given uid or gid is supposed to be mapped
to. This contrasts with the chown(2) syscall which cannot by itself
take the current ownership of the files it changes into account. It
simply changes the ownership to the specified uid and gid. This is
especially problematic when recursively chown(2)ing a large set of
files which is commong with the aforementioned portable home
directory and container and vm scenario.
- Idmapped mounts allow to change ownership locally, restricting it
to specific mounts, and temporarily as the ownership changes only
apply as long as the mount exists.
Several userspace projects have either already put up patches and
pull-requests for this feature or will do so should you decide to pull
this:
- systemd: In a wide variety of scenarios but especially right away
in their implementation of portable home directories.
https://systemd.io/HOME_DIRECTORY/
- container runtimes: containerd, runC, LXD:To share data between
host and unprivileged containers, unprivileged and privileged
containers, etc. The pull request for idmapped mounts support in
containerd, the default Kubernetes runtime is already up for quite
a while now: https://github.com/containerd/containerd/pull/4734
- The virtio-fs developers and several users have expressed interest
in using this feature with virtual machines once virtio-fs is
ported.
- ChromeOS: Sharing host-directories with unprivileged containers.
I've tightly synced with all those projects and all of those listed
here have also expressed their need/desire for this feature on the
mailing list. For more info on how people use this there's a bunch of
talks about this too. Here's just two recent ones:
https://www.cncf.io/wp-content/uploads/2020/12/Rootless-Containers-in-Gitpod.pdf
https://fosdem.org/2021/schedule/event/containers_idmap/
This comes with an extensive xfstests suite covering both ext4 and
xfs:
https://git.kernel.org/brauner/xfstests-dev/h/idmapped_mounts
It covers truncation, creation, opening, xattrs, vfscaps, setid
execution, setgid inheritance and more both with idmapped and
non-idmapped mounts. It already helped to discover an unrelated xfs
setgid inheritance bug which has since been fixed in mainline. It will
be sent for inclusion with the xfstests project should you decide to
merge this.
In order to support per-mount idmappings vfsmounts are marked with
user namespaces. The idmapping of the user namespace will be used to
map the ids of vfs objects when they are accessed through that mount.
By default all vfsmounts are marked with the initial user namespace.
The initial user namespace is used to indicate that a mount is not
idmapped. All operations behave as before and this is verified in the
testsuite.
Based on prior discussions we want to attach the whole user namespace
and not just a dedicated idmapping struct. This allows us to reuse all
the helpers that already exist for dealing with idmappings instead of
introducing a whole new range of helpers. In addition, if we decide in
the future that we are confident enough to enable unprivileged users
to setup idmapped mounts the permission checking can take into account
whether the caller is privileged in the user namespace the mount is
currently marked with.
The user namespace the mount will be marked with can be specified by
passing a file descriptor refering to the user namespace as an
argument to the new mount_setattr() syscall together with the new
MOUNT_ATTR_IDMAP flag. The system call follows the openat2() pattern
of extensibility.
The following conditions must be met in order to create an idmapped
mount:
- The caller must currently have the CAP_SYS_ADMIN capability in the
user namespace the underlying filesystem has been mounted in.
- The underlying filesystem must support idmapped mounts.
- The mount must not already be idmapped. This also implies that the
idmapping of a mount cannot be altered once it has been idmapped.
- The mount must be a detached/anonymous mount, i.e. it must have
been created by calling open_tree() with the OPEN_TREE_CLONE flag
and it must not already have been visible in the filesystem.
The last two points guarantee easier semantics for userspace and the
kernel and make the implementation significantly simpler.
By default vfsmounts are marked with the initial user namespace and no
behavioral or performance changes are observed.
The manpage with a detailed description can be found here:
https://git.kernel.org/brauner/man-pages/c/1d7b902e2875a1ff342e036a9f866a995640aea8
In order to support idmapped mounts, filesystems need to be changed
and mark themselves with the FS_ALLOW_IDMAP flag in fs_flags. The
patches to convert individual filesystem are not very large or
complicated overall as can be seen from the included fat, ext4, and
xfs ports. Patches for other filesystems are actively worked on and
will be sent out separately. The xfstestsuite can be used to verify
that port has been done correctly.
The mount_setattr() syscall is motivated independent of the idmapped
mounts patches and it's been around since July 2019. One of the most
valuable features of the new mount api is the ability to perform
mounts based on file descriptors only.
Together with the lookup restrictions available in the openat2()
RESOLVE_* flag namespace which we added in v5.6 this is the first time
we are close to hardened |
