]> rtime.felk.cvut.cz Git - linux-imx.git/commitdiff
exec: do not abuse ->cred_guard_mutex in threadgroup_lock()
authorOleg Nesterov <oleg@redhat.com>
Tue, 30 Apr 2013 22:28:20 +0000 (15:28 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Wed, 1 May 2013 00:04:07 +0000 (17:04 -0700)
threadgroup_lock() takes signal->cred_guard_mutex to ensure that
thread_group_leader() is stable.  This doesn't look nice, the scope of
this lock in do_execve() is huge.

And as Dave pointed out this can lead to deadlock, we have the
following dependencies:

do_execve: cred_guard_mutex -> i_mutex
cgroup_mount: i_mutex -> cgroup_mutex
attach_task_by_pid: cgroup_mutex -> cred_guard_mutex

Change de_thread() to take threadgroup_change_begin() around the
switch-the-leader code and change threadgroup_lock() to avoid
->cred_guard_mutex.

Note that de_thread() can't sleep with ->group_rwsem held, this can
obviously deadlock with the exiting leader if the writer is active, so it
does threadgroup_change_end() before schedule().

Reported-by: Dave Jones <davej@redhat.com>
Acked-by: Tejun Heo <tj@kernel.org>
Acked-by: Li Zefan <lizefan@huawei.com>
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
fs/exec.c
include/linux/sched.h

index 260f89f66651ec3124425a832dbf56bb8ea25df6..963f510a25ab50013e9b0fbeac055570dc8af235 100644 (file)
--- a/fs/exec.c
+++ b/fs/exec.c
@@ -898,11 +898,13 @@ static int de_thread(struct task_struct *tsk)
 
                sig->notify_count = -1; /* for exit_notify() */
                for (;;) {
+                       threadgroup_change_begin(tsk);
                        write_lock_irq(&tasklist_lock);
                        if (likely(leader->exit_state))
                                break;
                        __set_current_state(TASK_KILLABLE);
                        write_unlock_irq(&tasklist_lock);
+                       threadgroup_change_end(tsk);
                        schedule();
                        if (unlikely(__fatal_signal_pending(tsk)))
                                goto killed;
@@ -960,6 +962,7 @@ static int de_thread(struct task_struct *tsk)
                if (unlikely(leader->ptrace))
                        __wake_up_parent(leader, leader->parent);
                write_unlock_irq(&tasklist_lock);
+               threadgroup_change_end(tsk);
 
                release_task(leader);
        }
index a22baf83c20c17bbeedf9597379da02e44c9ccb3..6f950048b6e90fd8e59b2eb376377310a38cc407 100644 (file)
@@ -2249,27 +2249,18 @@ static inline void threadgroup_change_end(struct task_struct *tsk)
  *
  * Lock the threadgroup @tsk belongs to.  No new task is allowed to enter
  * and member tasks aren't allowed to exit (as indicated by PF_EXITING) or
- * perform exec.  This is useful for cases where the threadgroup needs to
- * stay stable across blockable operations.
+ * change ->group_leader/pid.  This is useful for cases where the threadgroup
+ * needs to stay stable across blockable operations.
  *
  * fork and exit paths explicitly call threadgroup_change_{begin|end}() for
  * synchronization.  While held, no new task will be added to threadgroup
  * and no existing live task will have its PF_EXITING set.
  *
- * During exec, a task goes and puts its thread group through unusual
- * changes.  After de-threading, exclusive access is assumed to resources
- * which are usually shared by tasks in the same group - e.g. sighand may
- * be replaced with a new one.  Also, the exec'ing task takes over group
- * leader role including its pid.  Exclude these changes while locked by
- * grabbing cred_guard_mutex which is used to synchronize exec path.
+ * de_thread() does threadgroup_change_{begin|end}() when a non-leader
+ * sub-thread becomes a new leader.
  */
 static inline void threadgroup_lock(struct task_struct *tsk)
 {
-       /*
-        * exec uses exit for de-threading nesting group_rwsem inside
-        * cred_guard_mutex. Grab cred_guard_mutex first.
-        */
-       mutex_lock(&tsk->signal->cred_guard_mutex);
        down_write(&tsk->signal->group_rwsem);
 }
 
@@ -2282,7 +2273,6 @@ static inline void threadgroup_lock(struct task_struct *tsk)
 static inline void threadgroup_unlock(struct task_struct *tsk)
 {
        up_write(&tsk->signal->group_rwsem);
-       mutex_unlock(&tsk->signal->cred_guard_mutex);
 }
 #else
 static inline void threadgroup_change_begin(struct task_struct *tsk) {}