[PATCH v5] Bluetooth: Keep MSFT ext info throughout a hci_dev's life cycle

[PATCH v5] Bluetooth: Keep MSFT ext info throughout a hci_dev's life cycle

[Manish Mandlik]

From: Miao-chen Chou <mcchou@chromium.org>

This moves msft_do_close() from hci_dev_do_close() to
hci_unregister_dev() to avoid clearing MSFT extension info. This also
re-reads MSFT info upon every msft_do_open() even if MSFT extension has
been initialized.

The following test steps were performed.
(1) boot the test device and verify the MSFT support debug log in syslog
(2) restart bluetoothd and verify msft_do_close() doesn't get invoked
    and msft_do_open re-reads the MSFT support.

Signed-off-by: Miao-chen Chou <mcchou@chromium.org>
Reviewed-by: Abhishek Pandit-Subedi <abhishekpandit@chromium.org>
Reviewed-by: Archie Pusaka <apusaka@chromium.org>
Reviewed-by: Alain Michaud <alainm@chromium.org>
Signed-off-by: Manish Mandlik <mmandlik@google.com>
---

Changes in v5:
- Rebase on ToT and remove extra blank line

Changes in v4:
- Re-read the MSFT data instead of skipping if it's initiated already

Changes in v3:
- Remove the accepted commits from the series

 net/bluetooth/hci_core.c |  3 ++-
 net/bluetooth/msft.c     | 20 +++++++++++++++++---
 2 files changed, 19 insertions(+), 4 deletions(-)

diff --git a/net/bluetooth/hci_core.c b/net/bluetooth/hci_core.c
index fb296478b86e..681c6dabb550 100644
--- a/net/bluetooth/hci_core.c
+++ b/net/bluetooth/hci_core.c
@@ -1798,7 +1798,6 @@ int hci_dev_do_close(struct hci_dev *hdev)
 	hci_sock_dev_event(hdev, HCI_DEV_DOWN);
 
 	aosp_do_close(hdev);
-	msft_do_close(hdev);
 
 	if (hdev->flush)
 		hdev->flush(hdev);
@@ -4026,6 +4025,8 @@ void hci_unregister_dev(struct hci_dev *hdev)
 		cancel_work_sync(&hdev->suspend_prepare);
 	}
 
+	msft_do_close(hdev);
+
 	hci_dev_do_close(hdev);
 
 	if (!test_bit(HCI_INIT, &hdev->flags) &&
diff --git a/net/bluetooth/msft.c b/net/bluetooth/msft.c
index b4bfae41e8a5..c2568e93598a 100644
--- a/net/bluetooth/msft.c
+++ b/net/bluetooth/msft.c
@@ -184,16 +184,30 @@ static void reregister_monitor_on_restart(struct hci_dev *hdev, int handle)
 
 void msft_do_open(struct hci_dev *hdev)
 {
-	struct msft_data *msft;
+	struct msft_data *msft = NULL;
 
 	if (hdev->msft_opcode == HCI_OP_NOP)
 		return;
 
 	bt_dev_dbg(hdev, "Initialize MSFT extension");
 
-	msft = kzalloc(sizeof(*msft), GFP_KERNEL);
-	if (!msft)
+	/* If MSFT data exists, reset its members */
+	if (hdev->msft_data) {
+		msft = hdev->msft_data;
+		hdev->msft_data = NULL;
+
+		msft->features = 0;
+		kfree(msft->evt_prefix);
+		msft->evt_prefix = NULL;
+		msft->evt_prefix_len = 0;
+	} else {
+		msft = kzalloc(sizeof(*msft), GFP_KERNEL);
+	}
+
+	if (!msft) {
+		bt_dev_err(hdev, "Failed to init MSFT extension");
 		return;
+	}
 
 	if (!read_supported_features(hdev, msft)) {
 		kfree(msft);
-- 
2.33.0.rc2.250.ged5fa647cd-goog

Re: [PATCH v5] Bluetooth: Keep MSFT ext info throughout a hci_dev's life cycle

[Marcel Holtmann]

Hi Manish,

> This moves msft_do_close() from hci_dev_do_close() to
> hci_unregister_dev() to avoid clearing MSFT extension info. This also
> re-reads MSFT info upon every msft_do_open() even if MSFT extension has
> been initialized.
> 
> The following test steps were performed.
> (1) boot the test device and verify the MSFT support debug log in syslog
> (2) restart bluetoothd and verify msft_do_close() doesn't get invoked
>    and msft_do_open re-reads the MSFT support.

so tell me how this can be correct. The msft_do_close does cleanup of instances. If we close the device via power down I would expect that these instances are cleared. Do they survive a HCI Reset command?

I think it would be better to introduce an additional msft_register / msft_unregister pair if this needs to be more complex.

Regards

Marcel