On Wed 03 Oct 04:34 PDT 2018, Arun Kumar Neelakantam wrote:
> Add support to wait on poll to get signal notifications.
> Send POLLPRI mask to indicate the signal change.
>
> Signed-off-by: Arun Kumar Neelakantam <aneela@xxxxxxxxxxxxxx>
> ---
> drivers/rpmsg/rpmsg_char.c | 14 ++++++++++++++
> 1 file changed, 14 insertions(+)
>
> diff --git a/drivers/rpmsg/rpmsg_char.c b/drivers/rpmsg/rpmsg_char.c
> index c983d6c..aee5561 100644
> --- a/drivers/rpmsg/rpmsg_char.c
> +++ b/drivers/rpmsg/rpmsg_char.c
> @@ -64,6 +64,7 @@ struct rpmsg_ctrldev {
> * @queue_lock: synchronization of @queue operations
> * @queue: incoming message queue
> * @readq: wait object for incoming queue
> + * @sig_pending:state of signal notification
> */
> struct rpmsg_eptdev {
> struct device dev;
> @@ -78,6 +79,8 @@ struct rpmsg_eptdev {
> spinlock_t queue_lock;
> struct sk_buff_head queue;
> wait_queue_head_t readq;
> +
> + atomic_t sig_pending;
> };
>
> static int rpmsg_eptdev_destroy(struct device *dev, void *data)
> @@ -125,6 +128,12 @@ static int rpmsg_ept_cb(struct rpmsg_device *rpdev, void *buf, int len,
> static int rpmsg_sigs_cb(struct rpmsg_device *rpdev, void *priv,
> u32 old, u32 new)
> {
> + struct rpmsg_eptdev *eptdev = priv;
> +
> + atomic_set(&eptdev->sig_pending, 1);
> +
> + /* wake up any blocking processes, waiting for signal notification */
> + wake_up_interruptible(&eptdev->readq);
This will execute a full memory barrier, so you shouldn't need to use
atomic_t to carry this information.
> return 0;
> }
>
The rest looks good.
Regards,
Bjorn
