[PATCH v3 1/2] Bluetooth: Send device found event on name resolve failure

[PATCH v3 1/2] Bluetooth: Send device found event on name resolve failure

[Archie Pusaka]

From: Archie Pusaka <apusaka@chromium.org>

Introducing NAME_REQUEST_FAILED flag that will be sent together with
device found event on name resolve failure. This will provide the
userspace with an information so it can decide not to resolve the
name for these devices in the future.

Signed-off-by: Archie Pusaka <apusaka@chromium.org>
Reviewed-by: Miao-chen Chou <mcchou@chromium.org>

---
Hi maintainers,

This is the patch series for remote name request as was discussed here.
https://patchwork.kernel.org/project/bluetooth/patch/20211028191805.1.I35b7f3a496f834de6b43a32f94b6160cb1467c94@changeid/
Please also review the corresponding userspace change.

Thanks,
Archie

Changes in v3:
* Reindent defines
* Assign variables inside if block instead of initializing

Changes in v2:
* Remove the part which accepts DONT_CARE flag in MGMT_OP_CONFIRM_NAME
* Rename MGMT constant to conform with the docs

 include/net/bluetooth/mgmt.h |  9 +++++----
 net/bluetooth/hci_event.c    | 11 ++++-------
 net/bluetooth/mgmt.c         | 12 ++++++++++--
 3 files changed, 19 insertions(+), 13 deletions(-)

diff --git a/include/net/bluetooth/mgmt.h b/include/net/bluetooth/mgmt.h
index 23a0524061b7..107b25deae68 100644
--- a/include/net/bluetooth/mgmt.h
+++ b/include/net/bluetooth/mgmt.h
@@ -936,10 +936,11 @@ struct mgmt_ev_auth_failed {
 	__u8	status;
 } __packed;
 
-#define MGMT_DEV_FOUND_CONFIRM_NAME    0x01
-#define MGMT_DEV_FOUND_LEGACY_PAIRING  0x02
-#define MGMT_DEV_FOUND_NOT_CONNECTABLE 0x04
-#define MGMT_DEV_FOUND_INITIATED_CONN  0x08
+#define MGMT_DEV_FOUND_CONFIRM_NAME		0x01
+#define MGMT_DEV_FOUND_LEGACY_PAIRING		0x02
+#define MGMT_DEV_FOUND_NOT_CONNECTABLE		0x04
+#define MGMT_DEV_FOUND_INITIATED_CONN		0x08
+#define MGMT_DEV_FOUND_NAME_REQUEST_FAILED	0x10
 
 #define MGMT_EV_DEVICE_FOUND		0x0012
 struct mgmt_ev_device_found {
diff --git a/net/bluetooth/hci_event.c b/net/bluetooth/hci_event.c
index dee4ef22fc88..bb4c04aecccf 100644
--- a/net/bluetooth/hci_event.c
+++ b/net/bluetooth/hci_event.c
@@ -2175,13 +2175,10 @@ static void hci_check_pending_name(struct hci_dev *hdev, struct hci_conn *conn,
 		return;
 
 	list_del(&e->list);
-	if (name) {
-		e->name_state = NAME_KNOWN;
-		mgmt_remote_name(hdev, bdaddr, ACL_LINK, 0x00,
-				 e->data.rssi, name, name_len);
-	} else {
-		e->name_state = NAME_NOT_KNOWN;
-	}
+
+	e->name_state = name ? NAME_KNOWN : NAME_NOT_KNOWN;
+	mgmt_remote_name(hdev, bdaddr, ACL_LINK, 0x00, e->data.rssi,
+			 name, name_len);
 
 	if (hci_resolve_next_name(hdev))
 		return;
diff --git a/net/bluetooth/mgmt.c b/net/bluetooth/mgmt.c
index f8f74d344297..bf989ae03f9f 100644
--- a/net/bluetooth/mgmt.c
+++ b/net/bluetooth/mgmt.c
@@ -9615,6 +9615,7 @@ void mgmt_remote_name(struct hci_dev *hdev, bdaddr_t *bdaddr, u8 link_type,
 	struct mgmt_ev_device_found *ev;
 	char buf[sizeof(*ev) + HCI_MAX_NAME_LENGTH + 2];
 	u16 eir_len;
+	u32 flags;
 
 	ev = (struct mgmt_ev_device_found *) buf;
 
@@ -9624,10 +9625,17 @@ void mgmt_remote_name(struct hci_dev *hdev, bdaddr_t *bdaddr, u8 link_type,
 	ev->addr.type = link_to_bdaddr(link_type, addr_type);
 	ev->rssi = rssi;
 
-	eir_len = eir_append_data(ev->eir, 0, EIR_NAME_COMPLETE, name,
-				  name_len);
+	if (name) {
+		eir_len = eir_append_data(ev->eir, 0, EIR_NAME_COMPLETE, name,
+					  name_len);
+		flags = 0;
+	} else {
+		eir_len = 0;
+		flags = MGMT_DEV_FOUND_NAME_REQUEST_FAILED;
+	}
 
 	ev->eir_len = cpu_to_le16(eir_len);
+	ev->flags = cpu_to_le32(flags);
 
 	mgmt_event(MGMT_EV_DEVICE_FOUND, hdev, ev, sizeof(*ev) + eir_len, NULL);
 }
-- 
2.34.0.rc2.393.gf8c9666880-goog

[PATCH v3 2/2] Bluetooth: Limit duration of Remote Name Resolve

[Archie Pusaka]

From: Archie Pusaka <apusaka@chromium.org>

When doing remote name request, we cannot scan. In the normal case it's
OK since we can expect it to finish within a short amount of time.
However, there is a possibility to scan lots of devices that
(1) requires Remote Name Resolve
(2) is unresponsive to Remote Name Resolve
When this happens, we are stuck to do Remote Name Resolve until all is
done before continue scanning.

This patch adds a time limit to stop us spending too long on remote
name request.

Signed-off-by: Archie Pusaka <apusaka@chromium.org>
Reviewed-by: Miao-chen Chou <mcchou@chromium.org>

---

Changes in v3:
* Add units in comment
* change debug log to warn

 include/net/bluetooth/hci_core.h | 3 +++
 net/bluetooth/hci_event.c        | 7 +++++++
 2 files changed, 10 insertions(+)

diff --git a/include/net/bluetooth/hci_core.h b/include/net/bluetooth/hci_core.h
index 2560cfe80db8..742cde106ae3 100644
--- a/include/net/bluetooth/hci_core.h
+++ b/include/net/bluetooth/hci_core.h
@@ -88,6 +88,7 @@ struct discovery_state {
 	u8			(*uuids)[16];
 	unsigned long		scan_start;
 	unsigned long		scan_duration;
+	unsigned long		name_resolve_timeout;
 };
 
 #define SUSPEND_NOTIFIER_TIMEOUT	msecs_to_jiffies(2000) /* 2 seconds */
@@ -1759,6 +1760,8 @@ void hci_mgmt_chan_unregister(struct hci_mgmt_chan *c);
 #define DISCOV_LE_FAST_ADV_INT_MIN	0x00A0	/* 100 msec */
 #define DISCOV_LE_FAST_ADV_INT_MAX	0x00F0	/* 150 msec */
 
+#define NAME_RESOLVE_DURATION		msecs_to_jiffies(10240)	/* 10.24 sec */
+
 void mgmt_fill_version_info(void *ver);
 int mgmt_new_settings(struct hci_dev *hdev);
 void mgmt_index_added(struct hci_dev *hdev);
diff --git a/net/bluetooth/hci_event.c b/net/bluetooth/hci_event.c
index bb4c04aecccf..779c70b21039 100644
--- a/net/bluetooth/hci_event.c
+++ b/net/bluetooth/hci_event.c
@@ -2129,6 +2129,12 @@ static bool hci_resolve_next_name(struct hci_dev *hdev)
 	if (list_empty(&discov->resolve))
 		return false;
 
+	/* We should stop if we already spent too much time resolving names. */
+	if (time_after(jiffies, discov->name_resolve_timeout)) {
+		bt_dev_warn_ratelimited(hdev, "Name resolve takes too long.");
+		return false;
+	}
+
 	e = hci_inquiry_cache_lookup_resolve(hdev, BDADDR_ANY, NAME_NEEDED);
 	if (!e)
 		return false;
@@ -2716,6 +2722,7 @@ static void hci_inquiry_complete_evt(struct hci_dev *hdev, struct sk_buff *skb)
 	if (e && hci_resolve_name(hdev, e) == 0) {
 		e->name_state = NAME_PENDING;
 		hci_discovery_set_state(hdev, DISCOVERY_RESOLVING);
+		discov->name_resolve_timeout = jiffies + NAME_RESOLVE_DURATION;
 	} else {
 		/* When BR/EDR inquiry is active and no LE scanning is in
 		 * progress, then change discovery state to indicate completion.
-- 
2.34.0.rc2.393.gf8c9666880-goog

Re: [PATCH v3 1/2] Bluetooth: Send device found event on name resolve failure

[Marcel Holtmann]

Hi Archie,

> Introducing NAME_REQUEST_FAILED flag that will be sent together with
> device found event on name resolve failure. This will provide the
> userspace with an information so it can decide not to resolve the
> name for these devices in the future.
> 
> Signed-off-by: Archie Pusaka <apusaka@chromium.org>
> Reviewed-by: Miao-chen Chou <mcchou@chromium.org>
> 
> ---
> Hi maintainers,
> 
> This is the patch series for remote name request as was discussed here.
> https://patchwork.kernel.org/project/bluetooth/patch/20211028191805.1.I35b7f3a496f834de6b43a32f94b6160cb1467c94@changeid/
> Please also review the corresponding userspace change.
> 
> Thanks,
> Archie
> 
> Changes in v3:
> * Reindent defines
> * Assign variables inside if block instead of initializing
> 
> Changes in v2:
> * Remove the part which accepts DONT_CARE flag in MGMT_OP_CONFIRM_NAME
> * Rename MGMT constant to conform with the docs
> 
> include/net/bluetooth/mgmt.h |  9 +++++----
> net/bluetooth/hci_event.c    | 11 ++++-------
> net/bluetooth/mgmt.c         | 12 ++++++++++--
> 3 files changed, 19 insertions(+), 13 deletions(-)

patch has been applied to bluetooth-next tree.

Regards

Marcel

Re: [PATCH v3 2/2] Bluetooth: Limit duration of Remote Name Resolve

[Marcel Holtmann]

Hi Archie,

> When doing remote name request, we cannot scan. In the normal case it's
> OK since we can expect it to finish within a short amount of time.
> However, there is a possibility to scan lots of devices that
> (1) requires Remote Name Resolve
> (2) is unresponsive to Remote Name Resolve
> When this happens, we are stuck to do Remote Name Resolve until all is
> done before continue scanning.
> 
> This patch adds a time limit to stop us spending too long on remote
> name request.
> 
> Signed-off-by: Archie Pusaka <apusaka@chromium.org>
> Reviewed-by: Miao-chen Chou <mcchou@chromium.org>
> 
> ---
> 
> Changes in v3:
> * Add units in comment
> * change debug log to warn
> 
> include/net/bluetooth/hci_core.h | 3 +++
> net/bluetooth/hci_event.c        | 7 +++++++
> 2 files changed, 10 insertions(+)

patch has been applied to bluetooth-next tree.

Regards

Marcel