exit: Factor coredump_exit_mm out of exit_mm

Separate the coredump logic from the ordinary exit_mm logic
by moving the coredump logic out of exit_mm into it's own
function coredump_exit_mm.

Link: https://lkml.kernel.org/r/87a6k2x277.fsf@disp2133
Reviewed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>

fs/coredump.c

@@ -404,8 +404,8 @@ static int zap_threads(struct task_struct *tsk, struct mm_struct *mm,
 	 *
 	 * do_exit:
 	 *	The caller holds mm->mmap_lock. This means that the task which
-	 *	uses this mm can't pass exit_mm(), so it can't exit or clear
+	 *	uses this mm can't pass coredump_exit_mm(), so it can't exit or
-	 *	its ->mm.
+	 *	clear its ->mm.
 	 *
 	 * de_thread:
 	 *	It does list_replace_rcu(&leader->tasks, &current->tasks),
@@ -500,7 +500,7 @@ static void coredump_finish(struct mm_struct *mm, bool core_dumped)
 		next = curr->next;
 		task = curr->task;
 		/*
-		 * see exit_mm(), curr->task must not see
+		 * see coredump_exit_mm(), curr->task must not see
 		 * ->task == NULL before we read ->next.
 		 */
 		smp_mb();

kernel/exit.c

@@ -339,6 +339,46 @@ kill_orphaned_pgrp(struct task_struct *tsk, struct task_struct *parent)
 	}
 }
 
+static void coredump_exit_mm(struct mm_struct *mm)
+{
+	struct core_state *core_state;
+
+	/*
+	 * Serialize with any possible pending coredump.
+	 * We must hold mmap_lock around checking core_state
+	 * and clearing tsk->mm.  The core-inducing thread
+	 * will increment ->nr_threads for each thread in the
+	 * group with ->mm != NULL.
+	 */
+	core_state = mm->core_state;
+	if (core_state) {
+		struct core_thread self;
+
+		mmap_read_unlock(mm);
+
+		self.task = current;
+		if (self.task->flags & PF_SIGNALED)
+			self.next = xchg(&core_state->dumper.next, &self);
+		else
+			self.task = NULL;
+		/*
+		 * Implies mb(), the result of xchg() must be visible
+		 * to core_state->dumper.
+		 */
+		if (atomic_dec_and_test(&core_state->nr_threads))
+			complete(&core_state->startup);
+
+		for (;;) {
+			set_current_state(TASK_UNINTERRUPTIBLE);
+			if (!self.task) /* see coredump_finish() */
+				break;
+			freezable_schedule();
+		}
+		__set_current_state(TASK_RUNNING);
+		mmap_read_lock(mm);
+	}
+}
+
 #ifdef CONFIG_MEMCG
 /*
  * A task is exiting.   If it owned this mm, find a new owner for the mm.
@@ -434,47 +474,13 @@ assign_new_owner:
 static void exit_mm(void)
 {
 	struct mm_struct *mm = current->mm;
-	struct core_state *core_state;
 
 	exit_mm_release(current, mm);
 	if (!mm)
 		return;
 	sync_mm_rss(mm);
-	/*
-	 * Serialize with any possible pending coredump.
-	 * We must hold mmap_lock around checking core_state
-	 * and clearing tsk->mm.  The core-inducing thread
-	 * will increment ->nr_threads for each thread in the
-	 * group with ->mm != NULL.
-	 */
 	mmap_read_lock(mm);
-	core_state = mm->core_state;
+	coredump_exit_mm(mm);
-	if (core_state) {
-		struct core_thread self;
-
-		mmap_read_unlock(mm);
-
-		self.task = current;
-		if (self.task->flags & PF_SIGNALED)
-			self.next = xchg(&core_state->dumper.next, &self);
-		else
-			self.task = NULL;
-		/*
-		 * Implies mb(), the result of xchg() must be visible
-		 * to core_state->dumper.
-		 */
-		if (atomic_dec_and_test(&core_state->nr_threads))
-			complete(&core_state->startup);
-
-		for (;;) {
-			set_current_state(TASK_UNINTERRUPTIBLE);
-			if (!self.task) /* see coredump_finish() */
-				break;
-			freezable_schedule();
-		}
-		__set_current_state(TASK_RUNNING);
-		mmap_read_lock(mm);
-	}
 	mmgrab(mm);
 	BUG_ON(mm != current->active_mm);
 	/* more a memory barrier than a real lock */

mm/oom_kill.c

@@ -787,9 +787,9 @@ static inline bool __task_will_free_mem(struct task_struct *task)
 	struct signal_struct *sig = task->signal;
 
 	/*
-	 * A coredumping process may sleep for an extended period in exit_mm(),
+	 * A coredumping process may sleep for an extended period in
-	 * so the oom killer cannot assume that the process will promptly exit
+	 * coredump_exit_mm(), so the oom killer cannot assume that
-	 * and release memory.
+	 * the process will promptly exit and release memory.
 	 */
 	if (sig->flags & SIGNAL_GROUP_COREDUMP)
 		return false;