Return-Path: <cyrus@holtmann.org>
Content-Type: text/plain; charset=us-ascii
Mime-Version: 1.0 (Mac OS X Mail 6.6 \(1510\))
Subject: Re: [RFC 15/15] Bluetooth: Auto connection and power off/on
From: Marcel Holtmann <marcel@holtmann.org>
In-Reply-To: <1381965485-9159-16-git-send-email-andre.guedes@openbossa.org>
Date: Thu, 17 Oct 2013 12:55:19 +0200
Cc: linux-bluetooth@vger.kernel.org
Message-Id: <2BA6ACEE-8390-48E0-883A-73738551FDDD@holtmann.org>
References: <1381965485-9159-1-git-send-email-andre.guedes@openbossa.org> <1381965485-9159-16-git-send-email-andre.guedes@openbossa.org>
To: Andre Guedes <andre.guedes@openbossa.org>
Sender: linux-bluetooth-owner@vger.kernel.org
List-ID: <linux-bluetooth.vger.kernel.org>

Hi Andre,

> If hdev is closed (e.g. Mgmt power off command, RFKILL or controller is
> reset), the established connections are dropped and no Disconnection
> Complete Event is sent to host. This way, the background scan is not
> triggered when devices configured with BT_AUTO_CONN_ALWAYS option
> disconnect. To fix this issue, before dropping the LE connections, we
> trigger the background scan for each connected device that requires
> BT_AUTO_CONN_ALWAYS auto connection.
> 
> Moreover, once the adapter is powered on, we should start the background
> scan if we have triggers registered. This way, we keep the background
> scan running after a power off and power on sequence.

these are actually two independent patches.

> Signed-off-by: Andre Guedes <andre.guedes@openbossa.org>
> ---
> net/bluetooth/hci_conn.c | 35 +++++++++++++++++++++++++++++++++++
> net/bluetooth/hci_core.c |  2 ++
> 2 files changed, 37 insertions(+)
> 
> diff --git a/net/bluetooth/hci_conn.c b/net/bluetooth/hci_conn.c
> index 5caf13b..66823eb 100644
> --- a/net/bluetooth/hci_conn.c
> +++ b/net/bluetooth/hci_conn.c
> @@ -954,6 +954,31 @@ timer:
> 				   msecs_to_jiffies(hdev->idle_timeout));
> }
> 
> +static void le_conn_drop_fixup(struct hci_conn *conn)
> +{
> +	struct hci_dev *hdev = conn->hdev;
> +	struct hci_conn_param *param;
> +	int err;
> +
> +	param = hci_find_conn_param(hdev, &conn->dst, conn->dst_type);
> +	if (!param)
> +		return;
> +
> +	if (param->auto_connect != BT_AUTO_CONN_ALWAYS)
> +		goto done;
> +
> +	err = hci_trigger_background_scan(hdev);
> +	if (err) {
> +		BT_ERR("Failed to trigger background scanning: %d", err);
> +		goto done;
> +	}
> +
> +	param->bg_scan_triggered = true;
> +
> +done:
> +	hci_conn_param_put(param);
> +}
> +
> /* Drop all connection on the device */
> void hci_conn_hash_flush(struct hci_dev *hdev)
> {
> @@ -963,6 +988,16 @@ void hci_conn_hash_flush(struct hci_dev *hdev)
> 	BT_DBG("hdev %s", hdev->name);
> 
> 	list_for_each_entry_safe(c, n, &h->list, list) {
> +		/* If this is a LE connection in connected state we should do
> +		 * some fixup before dropping this connection. Since no
> +		 * Disconnection Complete Event will be sent to the host, we
> +		 * have to trigger the background scan in case this is a
> +		 * BT_AUTO_CONN_ALWAYS device. This is handled by the le_conn_
> +		 * drop_fixup() helper.
> +		 */
> +		if (c->type == LE_LINK && c->state == BT_CONNECTED)
> +			le_conn_drop_fixup(c);
> +
> 		c->state = BT_CLOSED;

I do not like this part. We already have the case where we need to re-enable advertising after a connection drops. So this should be in a common place. I do not want to hack this in into all kinds of places.

Regards

Marcel