From: Andrea Parri (Microsoft) <[email protected]> Sent: Wednesday, 
November 18, 2020 6:37 AM
> 
> When channel->device_obj is non-NULL, vmbus_onoffer_rescind() could
> invoke put_device(), that will eventually release the device and free
> the channel object (cf. vmbus_device_release()).  However, a pointer
> to the object is dereferenced again later to load the primary_channel.
> The use-after-free can be avoided by noticing that this load/check is
> redundant if device_obk is non-NULL: primary_channel must be NULL if
> device_obj is non-NULL, cf. vmbus_add_channel_work().
> 
> Reported-by: Juan Vazquez <[email protected]>
> Signed-off-by: Andrea Parri (Microsoft) <[email protected]>
> ---
>  drivers/hv/channel_mgmt.c | 3 +--
>  1 file changed, 1 insertion(+), 2 deletions(-)
> 
> diff --git a/drivers/hv/channel_mgmt.c b/drivers/hv/channel_mgmt.c
> index 5bc5eef5da159..4072fd1f22146 100644
> --- a/drivers/hv/channel_mgmt.c
> +++ b/drivers/hv/channel_mgmt.c
> @@ -1116,8 +1116,7 @@ static void vmbus_onoffer_rescind(struct
> vmbus_channel_message_header *hdr)
>                       vmbus_device_unregister(channel->device_obj);
>                       put_device(dev);
>               }
> -     }
> -     if (channel->primary_channel != NULL) {
> +     } else if (channel->primary_channel != NULL) {
>               /*
>                * Sub-channel is being rescinded. Following is the channel
>                * close sequence when initiated from the driveri (refer to
> --
> 2.25.1

Taking into the account the separate comments from Wei Liu,

Reviewed-by: Michael Kelley <[email protected]>

Reply via email to