Message ID | 20190628123659.139576-4-sgarzare@redhat.com |
---|---|
State | Accepted |
Delegated to: | David Miller |
Headers | show |
Series | vsock/virtio: several fixes in the .probe() and .remove() | expand |
On Fri, Jun 28, 2019 at 02:36:59PM +0200, Stefano Garzarella wrote: > This patch moves the flush of works after vdev->config->del_vqs(vdev), > because we need to be sure that no workers run before to free the > 'vsock' object. > > Since we stopped the workers using the [tx|rx|event]_run flags, > we are sure no one is accessing the device while we are calling > vdev->config->reset(vdev), so we can safely move the workers' flush. What about send_pkt and loopback work? How were they stopped safely? For example, if send_pkt work executes then we're in trouble since it accesses the tx virtqueue which is deleted by ->del_vqs().
On Fri, Jun 28, 2019 at 02:36:59PM +0200, Stefano Garzarella wrote: > This patch moves the flush of works after vdev->config->del_vqs(vdev), > because we need to be sure that no workers run before to free the > 'vsock' object. > > Since we stopped the workers using the [tx|rx|event]_run flags, > we are sure no one is accessing the device while we are calling > vdev->config->reset(vdev), so we can safely move the workers' flush. > > Before the vdev->config->del_vqs(vdev), workers can be scheduled > by VQ callbacks, so we must flush them after del_vqs(), to avoid > use-after-free of 'vsock' object. Nevermind, I looked back at Patch 2 and saw the send_pkt and loopback work functions were also updated. Thanks! Stefan
diff --git a/net/vmw_vsock/virtio_transport.c b/net/vmw_vsock/virtio_transport.c index 1b44ec6f3f6c..96dafa978268 100644 --- a/net/vmw_vsock/virtio_transport.c +++ b/net/vmw_vsock/virtio_transport.c @@ -680,12 +680,6 @@ static void virtio_vsock_remove(struct virtio_device *vdev) rcu_assign_pointer(the_virtio_vsock, NULL); synchronize_rcu(); - flush_work(&vsock->loopback_work); - flush_work(&vsock->rx_work); - flush_work(&vsock->tx_work); - flush_work(&vsock->event_work); - flush_work(&vsock->send_pkt_work); - /* Reset all connected sockets when the device disappear */ vsock_for_each_connected_socket(virtio_vsock_reset_sock); @@ -740,6 +734,15 @@ static void virtio_vsock_remove(struct virtio_device *vdev) /* Delete virtqueues and flush outstanding callbacks if any */ vdev->config->del_vqs(vdev); + /* Other works can be queued before 'config->del_vqs()', so we flush + * all works before to free the vsock object to avoid use after free. + */ + flush_work(&vsock->loopback_work); + flush_work(&vsock->rx_work); + flush_work(&vsock->tx_work); + flush_work(&vsock->event_work); + flush_work(&vsock->send_pkt_work); + mutex_unlock(&the_virtio_vsock_mutex); kfree(vsock);
This patch moves the flush of works after vdev->config->del_vqs(vdev), because we need to be sure that no workers run before to free the 'vsock' object. Since we stopped the workers using the [tx|rx|event]_run flags, we are sure no one is accessing the device while we are calling vdev->config->reset(vdev), so we can safely move the workers' flush. Before the vdev->config->del_vqs(vdev), workers can be scheduled by VQ callbacks, so we must flush them after del_vqs(), to avoid use-after-free of 'vsock' object. Suggested-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Stefano Garzarella <sgarzare@redhat.com> --- net/vmw_vsock/virtio_transport.c | 15 +++++++++------ 1 file changed, 9 insertions(+), 6 deletions(-)