[PATCH] lsm: copy comm before calling audit_log to avoid race in string printing

From: Richard Guy Briggs
Date: Sun Nov 16 2014 - 16:45:00 EST

When task->comm is passed directly to audit_log_untrustedstring() without
getting a copy or using the task_lock, there is a race that could happen that
would output a NULL (\0) in the middle of the output string that would
effectively truncate the rest of the report text after the comm= field in the
audit log message, losing fields.

Using get_task_comm() to get a copy while acquiring the task_lock to prevent
this and to prevent the result from being a mixture of old and new values of
comm would incur potentially unacceptable overhead, considering that the value
can be influenced by userspace and therefore untrusted anyways.

Copy the value before passing it to audit_log_untrustedstring() ensures that a
local copy is used to calculate the length *and* subsequently printed. Even if
this value contains a mix of old and new values, it will only calculate and
copy up to the first NULL, preventing the rest of the audit log message being

The LSM_AUDIT_DATA_TASK pid= and comm= labels are duplicates of those at the
start of this function with different values. Rename them to their object
counterparts opid= and ocomm= to disambiguate. Use a second local copy of comm
to avoid a race between the first and second calls to
audit_log_untrustedstring() with comm.

Reported-by: Tetsuo Handa <penguin-kernel@xxxxxxxxxxxxxxxxxxx>
Signed-off-by: Richard Guy Briggs <rgb@xxxxxxxxxx>
security/lsm_audit.c | 17 ++++++++++-------
1 files changed, 10 insertions(+), 7 deletions(-)

diff --git a/security/lsm_audit.c b/security/lsm_audit.c
index 69fdf3b..3323144 100644
--- a/security/lsm_audit.c
+++ b/security/lsm_audit.c
@@ -211,7 +211,7 @@ static inline void print_ipv4_addr(struct audit_buffer *ab, __be32 addr,
static void dump_common_audit_data(struct audit_buffer *ab,
struct common_audit_data *a)
- struct task_struct *tsk = current;
+ char comm[sizeof(current->comm)];

* To keep stack sizes in check force programers to notice if they
@@ -220,8 +220,8 @@ static void dump_common_audit_data(struct audit_buffer *ab,
BUILD_BUG_ON(sizeof(a->u) > sizeof(void *)*2);

- audit_log_format(ab, " pid=%d comm=", task_pid_nr(tsk));
- audit_log_untrustedstring(ab, tsk->comm);
+ audit_log_format(ab, " pid=%d comm=", task_pid_nr(current));
+ audit_log_untrustedstring(ab, memcpy(comm, current->comm, sizeof(comm)));

switch (a->type) {
@@ -276,16 +276,19 @@ static void dump_common_audit_data(struct audit_buffer *ab,
audit_log_format(ab, " ino=%lu", inode->i_ino);
- tsk = a->u.tsk;
+ struct task_struct *tsk = a->u.tsk;
if (tsk) {
pid_t pid = task_pid_nr(tsk);
if (pid) {
- audit_log_format(ab, " pid=%d comm=", pid);
- audit_log_untrustedstring(ab, tsk->comm);
+ char comm[sizeof(tsk->comm)];
+ audit_log_format(ab, " opid=%d ocomm=", pid);
+ audit_log_untrustedstring(ab,
+ memcpy(comm, tsk->comm, sizeof(comm)));
+ }
if (a->u.net->sk) {
struct sock *sk = a->u.net->sk;

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/