[PATCH] Fix lguest oops when guest dies while receiving I/O

[Rusty Russell]

lguest needs to hold a reference to its task in case it exits while
another Guest is sending it I/O.  Otherwise we can oops in
access_process_vm->get_task_mm->task_lock().

Signed-off-by: Rusty Russell <rusty@xxxxxxxxxxxxxxx>
---
 drivers/lguest/lguest_user.c |    4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

===================================================================
--- a/drivers/lguest/lguest_user.c
+++ b/drivers/lguest/lguest_user.c
@@ -107,7 +107,8 @@ static int initialize(struct file *file,
 	setup_regs(lg->regs, args[2]);
 	setup_guest_gdt(lg);
 	lg->tsk = current;
-	lg->mm = get_task_mm(current);
+	get_task_struct(lg->tsk);
+	lg->mm = get_task_mm(lg->tsk);
 	lg->last_pages = NULL;
 	mutex_unlock(&lguest_lock);
 
@@ -160,6 +161,7 @@ static int close(struct inode *inode, st
 	mutex_lock(&lguest_lock);
 	release_all_dma(lg);
 	free_guest_pagetable(lg);
+	put_task_struct(lg->tsk);
 	mmput(lg->mm);
 	if (!IS_ERR(lg->dead))
 		kfree(lg->dead);


-
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/