4.9-stable review patch. If anyone has any objections, please let me know.
------------------
From: Hans Verkuil hverkuil@xs4all.nl
commit cd26d1c4d1bc947b56ae404998ae2276df7b39b7 upstream.
If a filehandle is dup()ped, then it is possible to close it from one fd and call mmap from the other. This creates a race condition in vb2_mmap where it is using queue data that __vb2_queue_free (called from close()) is in the process of releasing.
By moving up the mutex_lock(mmap_lock) in vb2_mmap this race is avoided since __vb2_queue_free is called with the same mutex locked. So vb2_mmap now reads consistent buffer data.
Signed-off-by: Hans Verkuil hverkuil@xs4all.nl Reported-by: syzbot+be93025dd45dccd8923c@syzkaller.appspotmail.com Signed-off-by: Hans Verkuil hansverk@cisco.com Signed-off-by: Mauro Carvalho Chehab mchehab+samsung@kernel.org Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org
--- drivers/media/v4l2-core/videobuf2-core.c | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-)
--- a/drivers/media/v4l2-core/videobuf2-core.c +++ b/drivers/media/v4l2-core/videobuf2-core.c @@ -1916,9 +1916,13 @@ int vb2_mmap(struct vb2_queue *q, struct return -EINVAL; } } + + mutex_lock(&q->mmap_lock); + if (vb2_fileio_is_active(q)) { dprintk(1, "mmap: file io in progress\n"); - return -EBUSY; + ret = -EBUSY; + goto unlock; }
/* @@ -1926,7 +1930,7 @@ int vb2_mmap(struct vb2_queue *q, struct */ ret = __find_plane_by_offset(q, off, &buffer, &plane); if (ret) - return ret; + goto unlock;
vb = q->bufs[buffer];
@@ -1942,8 +1946,9 @@ int vb2_mmap(struct vb2_queue *q, struct return -EINVAL; }
- mutex_lock(&q->mmap_lock); ret = call_memop(vb, mmap, vb->planes[plane].mem_priv, vma); + +unlock: mutex_unlock(&q->mmap_lock); if (ret) return ret;