[PATCH] lsm: get comm using lock to avoid race in string printing

[PATCH] lsm: get comm using lock to avoid race in string printing

[Richard Guy Briggs]

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 output string that would effectively truncate
the rest of the report text after the comm= field in the audit, losing fields.

Use 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.

Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: Richard Guy Briggs <rgb@redhat.com>
---
I've manually checked for locking issues and found none.  I've also enabled all the kernel lock debugging options and it came up clean.

 security/lsm_audit.c |    5 +++--
 1 files changed, 3 insertions(+), 2 deletions(-)

diff --git a/security/lsm_audit.c b/security/lsm_audit.c
index 69fdf3b..4773b91 100644
--- a/security/lsm_audit.c
+++ b/security/lsm_audit.c
@@ -212,6 +212,7 @@ static void dump_common_audit_data(struct audit_buffer *ab,
 				   struct common_audit_data *a)
 {
 	struct task_struct *tsk = current;
+	char comm[sizeof(tsk->comm)];
 
 	/*
 	 * To keep stack sizes in check force programers to notice if they
@@ -221,7 +222,7 @@ 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_untrustedstring(ab, get_task_comm(comm, tsk));
 
 	switch (a->type) {
 	case LSM_AUDIT_DATA_NONE:
@@ -282,7 +283,7 @@ static void dump_common_audit_data(struct audit_buffer *ab,
 			pid_t pid = task_pid_nr(tsk);
 			if (pid) {
 				audit_log_format(ab, " pid=%d comm=", pid);
-				audit_log_untrustedstring(ab, tsk->comm);
+				audit_log_untrustedstring(ab, get_task_comm(comm, tsk));
 			}
 		}
 		break;
-- 
1.7.1

Re: [PATCH] lsm: get comm using lock to avoid race in string printing

[Paul Moore]

On Friday, September 19, 2014 11:41:15 AM Richard Guy Briggs wrote:
> 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 output string that would effectively
> truncate the rest of the report text after the comm= field in the audit,
> losing fields.
> 
> Use 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.
> 
> Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
> Signed-off-by: Richard Guy Briggs <rgb@redhat.com>

The above is a bit odd ... is that a "Signed-off-by:" from Tetsuo Handa or a 
"From:"?

> ---
> I've manually checked for locking issues and found none.  I've also enabled
> all the kernel lock debugging options and it came up clean.
> 
>  security/lsm_audit.c |    5 +++--
>  1 files changed, 3 insertions(+), 2 deletions(-)
> 
> diff --git a/security/lsm_audit.c b/security/lsm_audit.c
> index 69fdf3b..4773b91 100644
> --- a/security/lsm_audit.c
> +++ b/security/lsm_audit.c
> @@ -212,6 +212,7 @@ static void dump_common_audit_data(struct audit_buffer
> *ab, struct common_audit_data *a)
>  {
>  	struct task_struct *tsk = current;
> +	char comm[sizeof(tsk->comm)];

This makes me a bit nervous for a potential race condition between allocation 
and use below.  How about using TASK_COMM_LEN instead?

>  	/*
>  	 * To keep stack sizes in check force programers to notice if they
> @@ -221,7 +222,7 @@ 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_untrustedstring(ab, get_task_comm(comm, tsk));
> 
>  	switch (a->type) {
>  	case LSM_AUDIT_DATA_NONE:
> @@ -282,7 +283,7 @@ static void dump_common_audit_data(struct audit_buffer
> *ab, pid_t pid = task_pid_nr(tsk);
>  			if (pid) {
>  				audit_log_format(ab, " pid=%d comm=", pid);
> -				audit_log_untrustedstring(ab, tsk->comm);
> +				audit_log_untrustedstring(ab, get_task_comm(comm, tsk));
>  			}
>  		}
>  		break;

-- 
paul moore
security and virtualization @ redhat

Re: [PATCH] lsm: get comm using lock to avoid race in string printing

[Richard Guy Briggs]

On 14/10/30, Paul Moore wrote:
> On Friday, September 19, 2014 11:41:15 AM Richard Guy Briggs wrote:
> > 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 output string that would effectively
> > truncate the rest of the report text after the comm= field in the audit,
> > losing fields.
> > 
> > Use 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.
> > 
> > Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
> > Signed-off-by: Richard Guy Briggs <rgb@redhat.com>
> 
> The above is a bit odd ... is that a "Signed-off-by:" from Tetsuo Handa or a 
> "From:"?

That was a "Signed-off-by:" that's likely my fault when manually
importing Tetsuo's oddly-formed patch email.

> > ---
> > I've manually checked for locking issues and found none.  I've also enabled
> > all the kernel lock debugging options and it came up clean.
> > 
> >  security/lsm_audit.c |    5 +++--
> >  1 files changed, 3 insertions(+), 2 deletions(-)
> > 
> > diff --git a/security/lsm_audit.c b/security/lsm_audit.c
> > index 69fdf3b..4773b91 100644
> > --- a/security/lsm_audit.c
> > +++ b/security/lsm_audit.c
> > @@ -212,6 +212,7 @@ static void dump_common_audit_data(struct audit_buffer
> > *ab, struct common_audit_data *a)
> >  {
> >  	struct task_struct *tsk = current;
> > +	char comm[sizeof(tsk->comm)];
> 
> This makes me a bit nervous for a potential race condition between allocation 
> and use below.  How about using TASK_COMM_LEN instead?

I don't have a strong opinion.  Both should work since tsk->comm is
defined as "char comm[TASK_COMM_LEN];" in struct task_struct.

> >  	/*
> >  	 * To keep stack sizes in check force programers to notice if they
> > @@ -221,7 +222,7 @@ 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_untrustedstring(ab, get_task_comm(comm, tsk));
> > 
> >  	switch (a->type) {
> >  	case LSM_AUDIT_DATA_NONE:
> > @@ -282,7 +283,7 @@ static void dump_common_audit_data(struct audit_buffer
> > *ab, pid_t pid = task_pid_nr(tsk);
> >  			if (pid) {
> >  				audit_log_format(ab, " pid=%d comm=", pid);
> > -				audit_log_untrustedstring(ab, tsk->comm);
> > +				audit_log_untrustedstring(ab, get_task_comm(comm, tsk));
> >  			}
> >  		}
> >  		break;
> 
> paul moore

- RGB

--
Richard Guy Briggs <rbriggs@redhat.com>
Senior Software Engineer, Kernel Security, AMER ENG Base Operating Systems, Red Hat
Remote, Ottawa, Canada
Voice: +1.647.777.2635, Internal: (81) 32635, Alt: +1.613.693.0684x3545