| Age | Commit message (Collapse) | Author |
|---|
|
AppArmor's begin_current_label_crit_section() is a scary function called
from lots of LSM hooks (in particular VFS/socket-related ones) that checks
if the label referenced by the current creds is marked FLAG_STALE, and if
so, attempts to use aa_replace_current_label() to replace the creds with an
updated version that uses a new label.
The first problem with this is that it would directly lead to UAF of
`struct cred` if anything in the kernel takes a pointer to the current
creds and accesses these past a security hook invocation that replaces
creds, like so:
```
const struct cred *cred = current_cred();
alloc_file_pseudo(...);
uid_t uid = cred->euid;
```
I don't know if anything in the kernel actually does this, but I think it
is very surprising that this pattern could lead to UAF.
The second problem is that things go wrong when aa_replace_current_label()
runs with overridden credentials. aa_replace_current_label() bails out if
`current_cred() != current_real_cred()` (mirroring the check in
proc_pid_attr_write()), but this check can't actually reliably detect
overridden credentials because the overridden creds can be the same as the
objective creds.
So in approximately the following scenario, things go wrong:
1. task begins with <creds A> (as both objective and subjective creds),
with refcount=2
2. task grabs an extra reference on <creds A> for overriding
3. task calls override_creds(<creds A>), which returns a pointer to the old
subjective creds (<creds A>)
4. task enters AppArmor LSM hook
5. AppArmor checks that objective/subjective creds are equal
6. AppArmor replaces both cred pointers with <creds B> and drops 2 refs on
<creds A>
7. task leaves AppArmor LSM hook
8. task calls revert_creds(<creds A>)
9. now task->cred is <creds A> while task->real_cred is <creds B>, but the
task_struct logically holds two references to <creds B>
10. another task drops the extra reference on <creds A> that was used for
overriding, refcount drops to 0
11. now task->real_cred points to freed creds
At this point, any access to current_cred() will be UAF.
I have a test case where I run aa-disable on a profile while a process
using that profile is blocked on splice() from a FUSE passthrough file into
a full pipe; after the profile update, the pipe becomes empty, splice()
resumes, the credentials go out of sync, and a subsequent getuid() syscall
results in a KASAN UAF splat.
To fix this, instead of directly replacing creds, do it via task_work that
will run at the end of the current syscall. (The point in time at which the
cred replacement happens should have no correctness impact; it is just a
performance optimization to avoid unnecessarily touching the refcount of
the new label.)
Note that AppArmor still performs direct cred replacements in the
sb_pivotroot LSM hook after this change, and that direct cred replacements
can still happen in VFS ->write() callbacks via proc_pid_attr_write().
There are two options for what to do with aa_dup_task_ctx(): Either
explicitly reset new->label_replacement_pending after the entire
aa_task_ctx has been copied, or switch to manually copying members over.
I am switching to manually copying members over because that should make
bugs more obvious.
Cc: stable@vger.kernel.org
Fixes: c75afcd153f6 ("AppArmor: contexts used in attaching policy to system objects")
Signed-off-by: Jann Horn <jannh@google.com>
Signed-off-by: John Johansen <john.johansen@canonical.com>
|
|
There is no need for a goto a label immediately following the
conditional block when the jump is the last statement in the block.
Fixes: 7306c41672487 ("apparmor: release exe file resources on path failure")
Signed-off-by: John Johansen <john.johansen@canonical.com>
|
|
get_current_exe_path() takes both an exe_file reference and a path
reference before resolving the path name. If aa_path_name() failed, it
returned immediately and leaked both references.
Route the failure through the common cleanup path so fput() and path_put()
always run after the references are acquired.
Fixes: 8d34e16f7f2b ("apparmor: userns: Add support for execpath in userns")
Reviewed-by: Ryan Lee <ryan.lee@canonical.com>
Signed-off-by: Zygmunt Krynicki <me@zygoon.pl>
Signed-off-by: John Johansen <john.johansen@canonical.com>
|
|
This new field allows reliable identification of the binary that
triggered a denial since the existing field (comm) only gives the name of
the binary, not its path. Thus comm doesn't work for binaries outside of
$PATH or works unreliably when two binaries have the same name.
Additionally comm can be modified by a program, for example, comm="(tor)"
or comm=4143504920506F6C6C6572 (= ACPI Poller).
Signed-off-by: Maxime Bélair <maxime.belair@canonical.com>
Signed-off-by: John Johansen <john.johansen@canonical.com>
|
|
The set of rules on a profile is not dynamically extended, instead
if a new ruleset is needed a new version of the profile is created.
This allows us to use a vector of rules instead of a list, slightly
reducing memory usage and simplifying the code.
Signed-off-by: John Johansen <john.johansen@canonical.com>
|
|
When the cred was explicit passed through to aa_may_ptrace() the
kernel-doc comment was not properly updated.
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202311040508.AUhi04RY-lkp@intel.com/
Signed-off-by: John Johansen <john.johansen@canonical.com>
|
|
Unprivileged user namespace creation is often used as a first step
in privilege escalation attacks. Instead of disabling it at the
sysrq level, which blocks its legitimate use as for setting up a sandbox,
allow control on a per domain basis.
This allows an admin to quickly lock down a system while also still
allowing legitimate use.
Reviewed-by: Georgia Garcia <georgia.garcia@canonical.com>
Signed-off-by: John Johansen <john.johansen@canonical.com>
|
|
The cred is needed to properly audit some messages, and will be needed
in the future for uid conditional mediation. So pass it through to
where the apparmor_audit_data struct gets defined.
Reviewed-by: Georgia Garcia <georgia.garcia@canonical.com>
Signed-off-by: John Johansen <john.johansen@canonical.com>
|
|
rename audit_data's label field to subj_label to better reflect its
use. Also at the same time drop unneeded assignments to ->subj_label
as the later call to aa_check_perms will do the assignment if needed.
Reviewed-by: Georgia Garcia <georgia.garcia@canonical.com>
Signed-off-by: John Johansen <john.johansen@canonical.com>
|
|
Everywhere where common_audit_data is used apparmor audit_data is also
used. We can simplify the code and drop the use of the aad macro
everywhere by combining the two structures.
Reviewed-by: Georgia Garcia <georgia.garcia@canonical.com>
Signed-off-by: John Johansen <john.johansen@canonical.com>
|
|
Change the return type to void since it always return 0, and no need
to do the checking in aa_set_current_onexec.
Signed-off-by: Quanfa Fu <quanfafu@gmail.com>
Reviewed-by: "Tyler Hicks (Microsoft)" <code@tyhicks.com>
Signed-off-by: John Johansen <john.johansen@canonical.com>
|
|
In aa_get_task_label(), aa_get_newest_cred_label(__task_cred(task))
can do the same things as aa_get_newest_label(__aa_task_raw_label(task)),
so we can replace it and remove __aa_task_raw_label() to simplify the code.
Signed-off-by: Gaosheng Cui <cuigaosheng1@huawei.com>
Signed-off-by: John Johansen <john.johansen@canonical.com>
|
|
Convert profile->rules to a list as the next step towards supporting
multiple rulesets in a profile. For this step only support a single
list entry item. The logic for iterating the list will come as a
separate step.
Signed-off-by: John Johansen <john.johansen@canonical.com>
|
|
In preparation for moving from a single set of rules and a single
attachment to multiple rulesets and attachments separate from the
profile refactor attachment information and ruleset info into their
own structures.
Signed-off-by: John Johansen <john.johansen@canonical.com>
|
|
Audit messages currently don't contain the mediation class which can
make them less clear than they should be in some circumstances. With
newer mediation classes coming this potential confusion will become
worse.
Fix this by adding the mediatin class to the messages.
Signed-off-by: John Johansen <john.johansen@canonical.com>
|
|
AppArmor split out task oriented controls to their own logical file
a while ago. Ptrace mediation is better grouped with task than
ipc, so move it.
Signed-off-by: John Johansen <john.johansen@canonical.com>
|
|
Based on 1 normalized pattern(s):
this program is free software you can redistribute it and or modify
it under the terms of the gnu general public license as published by
the free software foundation version 2 of the license
extracted by the scancode license scanner the SPDX license identifier
GPL-2.0-only
has been chosen to replace the boilerplate/reference in 315 file(s).
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Allison Randal <allison@lohutok.net>
Reviewed-by: Armijn Hemel <armijn@tjaldur.nl>
Cc: linux-spdx@vger.kernel.org
Link: https://lkml.kernel.org/r/20190531190115.503150771@linutronix.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
Don't use the cred->security pointer directly.
Provide a helper function that provides the security blob pointer.
Signed-off-by: Casey Schaufler <casey@schaufler-ca.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
[kees: adjusted for ordered init series]
Signed-off-by: Kees Cook <keescook@chromium.org>
|
|
Domain transition so far have been largely blocked by no new privs,
unless the transition has been provably a subset of the previous
confinement. There was a couple problems with the previous
implementations,
- transitions that weren't explicitly a stack but resulted in a subset
of confinement were disallowed
- confinement subsets were only calculated from the previous
confinement instead of the confinement being enforced at the time of
no new privs, so transitions would have to get progressively
tighter.
Fix this by detecting and storing a reference to the task's
confinement at the "time" no new privs is set. This reference is then
used to determine whether a transition is a subsystem of the
confinement at the time no new privs was set.
Unfortunately the implementation is less than ideal in that we have to
detect no new privs after the fact when a task attempts a domain
transition. This is adequate for the currently but will not work in a
stacking situation where no new privs could be conceivably be set in
both the "host" and in the container.
Signed-off-by: John Johansen <john.johansen@canonical.com>
|
|
Now that file contexts have been moved into file, and task context
fns() and data have been split from the context, only the cred context
remains in context.h so rename to cred.h to better reflect what it
deals with.
Signed-off-by: John Johansen <john.johansen@canonical.com>
|
|
Signed-off-by: John Johansen <john.johansen@canonical.com>
|
