From: Sean Wang <[email protected]>
There is a conflict between MediaTek wmt event and msft vendor extension
logic in the core layer since 145373cb1b1f ("Bluetooth: Add framework for
Microsoft vendor extension") was introduced because we changed the type of
mediatek wmt event to the type of msft vendor event in the driver.
But the purpose we reported mediatek event to the core layer is for the
diagnostic purpose with that we are able to see the full packet trace via
monitoring socket with btmon. Thus, it is harmless we keep the original
type of mediatek vendor event here to avoid breaking the msft extension
function especially they can be supported by Mediatek chipset like MT7921
, MT7922 devices and future devices.
Signed-off-by: Sean Wang <[email protected]>
---
drivers/bluetooth/btmtk.h | 1 +
drivers/bluetooth/btmtksdio.c | 9 +--------
drivers/bluetooth/btusb.c | 5 -----
3 files changed, 2 insertions(+), 13 deletions(-)
diff --git a/drivers/bluetooth/btmtk.h b/drivers/bluetooth/btmtk.h
index 8960a5f89d48..013850fd2055 100644
--- a/drivers/bluetooth/btmtk.h
+++ b/drivers/bluetooth/btmtk.h
@@ -5,6 +5,7 @@
#define FIRMWARE_MT7668 "mediatek/mt7668pr2h.bin"
#define FIRMWARE_MT7961 "mediatek/BT_RAM_CODE_MT7961_1_2_hdr.bin"
+#define HCI_EV_WMT 0xe4
#define HCI_WMT_MAX_EVENT_SIZE 64
#define BTMTK_WMT_REG_WRITE 0x1
diff --git a/drivers/bluetooth/btmtksdio.c b/drivers/bluetooth/btmtksdio.c
index 8e200e80d2f6..cbb09e1b823d 100644
--- a/drivers/bluetooth/btmtksdio.c
+++ b/drivers/bluetooth/btmtksdio.c
@@ -381,13 +381,6 @@ static int btmtksdio_recv_event(struct hci_dev *hdev, struct sk_buff *skb)
struct hci_event_hdr *hdr = (void *)skb->data;
int err;
- /* Fix up the vendor event id with 0xff for vendor specific instead
- * of 0xe4 so that event send via monitoring socket can be parsed
- * properly.
- */
- if (hdr->evt == 0xe4)
- hdr->evt = HCI_EV_VENDOR;
-
/* When someone waits for the WMT event, the skb is being cloned
* and being processed the events from there then.
*/
@@ -403,7 +396,7 @@ static int btmtksdio_recv_event(struct hci_dev *hdev, struct sk_buff *skb)
if (err < 0)
goto err_free_skb;
- if (hdr->evt == HCI_EV_VENDOR) {
+ if (hdr->evt == HCI_EV_WMT) {
if (test_and_clear_bit(BTMTKSDIO_TX_WAIT_VND_EVT,
&bdev->tx_state)) {
/* Barrier to sync with other CPUs */
diff --git a/drivers/bluetooth/btusb.c b/drivers/bluetooth/btusb.c
index aefa0ee293f3..55ebab019d29 100644
--- a/drivers/bluetooth/btusb.c
+++ b/drivers/bluetooth/btusb.c
@@ -2271,11 +2271,6 @@ static void btusb_mtk_wmt_recv(struct urb *urb)
skb_put_data(skb, urb->transfer_buffer, urb->actual_length);
hdr = (void *)skb->data;
- /* Fix up the vendor event id with 0xff for vendor specific
- * instead of 0xe4 so that event send via monitoring socket can
- * be parsed properly.
- */
- hdr->evt = 0xff;
/* When someone waits for the WMT event, the skb is being cloned
* and being processed the events from there then.
--
2.25.1
Hi Sean,
> There is a conflict between MediaTek wmt event and msft vendor extension
> logic in the core layer since 145373cb1b1f ("Bluetooth: Add framework for
> Microsoft vendor extension") was introduced because we changed the type of
> mediatek wmt event to the type of msft vendor event in the driver.
>
> But the purpose we reported mediatek event to the core layer is for the
> diagnostic purpose with that we are able to see the full packet trace via
> monitoring socket with btmon. Thus, it is harmless we keep the original
> type of mediatek vendor event here to avoid breaking the msft extension
> function especially they can be supported by Mediatek chipset like MT7921
> , MT7922 devices and future devices.
>
> Signed-off-by: Sean Wang <[email protected]>
> ---
> drivers/bluetooth/btmtk.h | 1 +
> drivers/bluetooth/btmtksdio.c | 9 +--------
> drivers/bluetooth/btusb.c | 5 -----
> 3 files changed, 2 insertions(+), 13 deletions(-)
>
> diff --git a/drivers/bluetooth/btmtk.h b/drivers/bluetooth/btmtk.h
> index 8960a5f89d48..013850fd2055 100644
> --- a/drivers/bluetooth/btmtk.h
> +++ b/drivers/bluetooth/btmtk.h
> @@ -5,6 +5,7 @@
> #define FIRMWARE_MT7668 "mediatek/mt7668pr2h.bin"
> #define FIRMWARE_MT7961 "mediatek/BT_RAM_CODE_MT7961_1_2_hdr.bin"
>
> +#define HCI_EV_WMT 0xe4
> #define HCI_WMT_MAX_EVENT_SIZE 64
>
> #define BTMTK_WMT_REG_WRITE 0x1
> diff --git a/drivers/bluetooth/btmtksdio.c b/drivers/bluetooth/btmtksdio.c
> index 8e200e80d2f6..cbb09e1b823d 100644
> --- a/drivers/bluetooth/btmtksdio.c
> +++ b/drivers/bluetooth/btmtksdio.c
> @@ -381,13 +381,6 @@ static int btmtksdio_recv_event(struct hci_dev *hdev, struct sk_buff *skb)
> struct hci_event_hdr *hdr = (void *)skb->data;
> int err;
>
> - /* Fix up the vendor event id with 0xff for vendor specific instead
> - * of 0xe4 so that event send via monitoring socket can be parsed
> - * properly.
> - */
> - if (hdr->evt == 0xe4)
> - hdr->evt = HCI_EV_VENDOR;
> -
> /* When someone waits for the WMT event, the skb is being cloned
> * and being processed the events from there then.
> */
> @@ -403,7 +396,7 @@ static int btmtksdio_recv_event(struct hci_dev *hdev, struct sk_buff *skb)
> if (err < 0)
> goto err_free_skb;
>
> - if (hdr->evt == HCI_EV_VENDOR) {
> + if (hdr->evt == HCI_EV_WMT) {
> if (test_and_clear_bit(BTMTKSDIO_TX_WAIT_VND_EVT,
> &bdev->tx_state)) {
> /* Barrier to sync with other CPUs */
> diff --git a/drivers/bluetooth/btusb.c b/drivers/bluetooth/btusb.c
> index aefa0ee293f3..55ebab019d29 100644
> --- a/drivers/bluetooth/btusb.c
> +++ b/drivers/bluetooth/btusb.c
> @@ -2271,11 +2271,6 @@ static void btusb_mtk_wmt_recv(struct urb *urb)
> skb_put_data(skb, urb->transfer_buffer, urb->actual_length);
>
> hdr = (void *)skb->data;
> - /* Fix up the vendor event id with 0xff for vendor specific
> - * instead of 0xe4 so that event send via monitoring socket can
> - * be parsed properly.
> - */
> - hdr->evt = 0xff;
>
> /* When someone waits for the WMT event, the skb is being cloned
> * and being processed the events from there then.
please fix the reported issue by kernel test robot.
Regards
Marcel
Hi Sean,
Thanks for your patch! Perhaps something to improve:
[auto build test WARNING on bluetooth-next/master]
[also build test WARNING on next-20220211]
[cannot apply to v5.17-rc3]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch]
url: https://github.com/0day-ci/linux/commits/sean-wang-mediatek-com/Bluetooth-mt7921s-support-bluetooth-reset-mechanism/20220129-140313
base: https://git.kernel.org/pub/scm/linux/kernel/git/bluetooth/bluetooth-next.git master
config: riscv-randconfig-c006-20220205 (https://download.01.org/0day-ci/archive/20220213/[email protected]/config)
compiler: clang version 15.0.0 (https://github.com/llvm/llvm-project dee058c670593b999fec19c458dbbd882ad9de56)
reproduce (this is a W=1 build):
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
chmod +x ~/bin/make.cross
# install riscv cross compiling tool for clang build
# apt-get install binutils-riscv64-linux-gnu
# https://github.com/0day-ci/linux/commit/33092fd43b74594bae07555a6fdf15133cc8ec54
git remote add linux-review https://github.com/0day-ci/linux
git fetch --no-tags linux-review sean-wang-mediatek-com/Bluetooth-mt7921s-support-bluetooth-reset-mechanism/20220129-140313
git checkout 33092fd43b74594bae07555a6fdf15133cc8ec54
# save the config file to linux build tree
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=clang make.cross ARCH=riscv clang-analyzer
If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <[email protected]>
clang-analyzer warnings: (new ones prefixed by >>)
>> drivers/bluetooth/btusb.c:2273:3: warning: Value stored to 'hdr' is never read [clang-analyzer-deadcode.DeadStores]
hdr = (void *)skb->data;
^ ~~~~~~~~~~~~~~~~~
vim +/hdr +2273 drivers/bluetooth/btusb.c
5a87679ffd4436 mark-yw.chen 2021-09-01 2248
a1c49c434e1505 Sean Wang 2019-06-02 2249 static void btusb_mtk_wmt_recv(struct urb *urb)
a1c49c434e1505 Sean Wang 2019-06-02 2250 {
a1c49c434e1505 Sean Wang 2019-06-02 2251 struct hci_dev *hdev = urb->context;
a1c49c434e1505 Sean Wang 2019-06-02 2252 struct btusb_data *data = hci_get_drvdata(hdev);
a1c49c434e1505 Sean Wang 2019-06-02 @2253 struct hci_event_hdr *hdr;
a1c49c434e1505 Sean Wang 2019-06-02 2254 struct sk_buff *skb;
a1c49c434e1505 Sean Wang 2019-06-02 2255 int err;
a1c49c434e1505 Sean Wang 2019-06-02 2256
a1c49c434e1505 Sean Wang 2019-06-02 2257 if (urb->status == 0 && urb->actual_length > 0) {
a1c49c434e1505 Sean Wang 2019-06-02 2258 hdev->stat.byte_rx += urb->actual_length;
a1c49c434e1505 Sean Wang 2019-06-02 2259
a1c49c434e1505 Sean Wang 2019-06-02 2260 /* WMT event shouldn't be fragmented and the size should be
a1c49c434e1505 Sean Wang 2019-06-02 2261 * less than HCI_WMT_MAX_EVENT_SIZE.
a1c49c434e1505 Sean Wang 2019-06-02 2262 */
a1c49c434e1505 Sean Wang 2019-06-02 2263 skb = bt_skb_alloc(HCI_WMT_MAX_EVENT_SIZE, GFP_ATOMIC);
a1c49c434e1505 Sean Wang 2019-06-02 2264 if (!skb) {
a1c49c434e1505 Sean Wang 2019-06-02 2265 hdev->stat.err_rx++;
60c6a63a3d3080 Mark-YW.Chen 2021-10-14 2266 kfree(urb->setup_packet);
de71a6cb4bf24d Jupeng Zhong 2021-02-02 2267 return;
a1c49c434e1505 Sean Wang 2019-06-02 2268 }
a1c49c434e1505 Sean Wang 2019-06-02 2269
a1c49c434e1505 Sean Wang 2019-06-02 2270 hci_skb_pkt_type(skb) = HCI_EVENT_PKT;
a1c49c434e1505 Sean Wang 2019-06-02 2271 skb_put_data(skb, urb->transfer_buffer, urb->actual_length);
a1c49c434e1505 Sean Wang 2019-06-02 2272
a1c49c434e1505 Sean Wang 2019-06-02 @2273 hdr = (void *)skb->data;
a1c49c434e1505 Sean Wang 2019-06-02 2274
a1c49c434e1505 Sean Wang 2019-06-02 2275 /* When someone waits for the WMT event, the skb is being cloned
a1c49c434e1505 Sean Wang 2019-06-02 2276 * and being processed the events from there then.
a1c49c434e1505 Sean Wang 2019-06-02 2277 */
a1c49c434e1505 Sean Wang 2019-06-02 2278 if (test_bit(BTUSB_TX_WAIT_VND_EVT, &data->flags)) {
22cc6b7a1dbb58 Johan Hovold 2019-11-28 2279 data->evt_skb = skb_clone(skb, GFP_ATOMIC);
de71a6cb4bf24d Jupeng Zhong 2021-02-02 2280 if (!data->evt_skb) {
de71a6cb4bf24d Jupeng Zhong 2021-02-02 2281 kfree_skb(skb);
60c6a63a3d3080 Mark-YW.Chen 2021-10-14 2282 kfree(urb->setup_packet);
de71a6cb4bf24d Jupeng Zhong 2021-02-02 2283 return;
de71a6cb4bf24d Jupeng Zhong 2021-02-02 2284 }
a1c49c434e1505 Sean Wang 2019-06-02 2285 }
a1c49c434e1505 Sean Wang 2019-06-02 2286
a1c49c434e1505 Sean Wang 2019-06-02 2287 err = hci_recv_frame(hdev, skb);
de71a6cb4bf24d Jupeng Zhong 2021-02-02 2288 if (err < 0) {
de71a6cb4bf24d Jupeng Zhong 2021-02-02 2289 kfree_skb(data->evt_skb);
de71a6cb4bf24d Jupeng Zhong 2021-02-02 2290 data->evt_skb = NULL;
60c6a63a3d3080 Mark-YW.Chen 2021-10-14 2291 kfree(urb->setup_packet);
de71a6cb4bf24d Jupeng Zhong 2021-02-02 2292 return;
de71a6cb4bf24d Jupeng Zhong 2021-02-02 2293 }
a1c49c434e1505 Sean Wang 2019-06-02 2294
a1c49c434e1505 Sean Wang 2019-06-02 2295 if (test_and_clear_bit(BTUSB_TX_WAIT_VND_EVT,
a1c49c434e1505 Sean Wang 2019-06-02 2296 &data->flags)) {
a1c49c434e1505 Sean Wang 2019-06-02 2297 /* Barrier to sync with other CPUs */
a1c49c434e1505 Sean Wang 2019-06-02 2298 smp_mb__after_atomic();
a1c49c434e1505 Sean Wang 2019-06-02 2299 wake_up_bit(&data->flags,
a1c49c434e1505 Sean Wang 2019-06-02 2300 BTUSB_TX_WAIT_VND_EVT);
a1c49c434e1505 Sean Wang 2019-06-02 2301 }
60c6a63a3d3080 Mark-YW.Chen 2021-10-14 2302 kfree(urb->setup_packet);
a1c49c434e1505 Sean Wang 2019-06-02 2303 return;
a1c49c434e1505 Sean Wang 2019-06-02 2304 } else if (urb->status == -ENOENT) {
a1c49c434e1505 Sean Wang 2019-06-02 2305 /* Avoid suspend failed when usb_kill_urb */
a1c49c434e1505 Sean Wang 2019-06-02 2306 return;
a1c49c434e1505 Sean Wang 2019-06-02 2307 }
a1c49c434e1505 Sean Wang 2019-06-02 2308
a1c49c434e1505 Sean Wang 2019-06-02 2309 usb_mark_last_busy(data->udev);
a1c49c434e1505 Sean Wang 2019-06-02 2310
a1c49c434e1505 Sean Wang 2019-06-02 2311 /* The URB complete handler is still called with urb->actual_length = 0
a1c49c434e1505 Sean Wang 2019-06-02 2312 * when the event is not available, so we should keep re-submitting
a1c49c434e1505 Sean Wang 2019-06-02 2313 * URB until WMT event returns, Also, It's necessary to wait some time
a1c49c434e1505 Sean Wang 2019-06-02 2314 * between the two consecutive control URBs to relax the target device
a1c49c434e1505 Sean Wang 2019-06-02 2315 * to generate the event. Otherwise, the WMT event cannot return from
a1c49c434e1505 Sean Wang 2019-06-02 2316 * the device successfully.
a1c49c434e1505 Sean Wang 2019-06-02 2317 */
48c13301e6baba Mark Chen 2021-02-02 2318 udelay(500);
a1c49c434e1505 Sean Wang 2019-06-02 2319
a1c49c434e1505 Sean Wang 2019-06-02 2320 usb_anchor_urb(urb, &data->ctrl_anchor);
a1c49c434e1505 Sean Wang 2019-06-02 2321 err = usb_submit_urb(urb, GFP_ATOMIC);
a1c49c434e1505 Sean Wang 2019-06-02 2322 if (err < 0) {
60c6a63a3d3080 Mark-YW.Chen 2021-10-14 2323 kfree(urb->setup_packet);
a1c49c434e1505 Sean Wang 2019-06-02 2324 /* -EPERM: urb is being killed;
a1c49c434e1505 Sean Wang 2019-06-02 2325 * -ENODEV: device got disconnected
a1c49c434e1505 Sean Wang 2019-06-02 2326 */
a1c49c434e1505 Sean Wang 2019-06-02 2327 if (err != -EPERM && err != -ENODEV)
a1c49c434e1505 Sean Wang 2019-06-02 2328 bt_dev_err(hdev, "urb %p failed to resubmit (%d)",
a1c49c434e1505 Sean Wang 2019-06-02 2329 urb, -err);
a1c49c434e1505 Sean Wang 2019-06-02 2330 usb_unanchor_urb(urb);
a1c49c434e1505 Sean Wang 2019-06-02 2331 }
a1c49c434e1505 Sean Wang 2019-06-02 2332 }
a1c49c434e1505 Sean Wang 2019-06-02 2333
---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/[email protected]