Re: [PATCH v2] bluetooth: hci_event: don't print an error on vendor events

[Luiz Augusto von Dentz]

Hi Marcel, Caleb,

On Wed, Mar 2, 2022 at 11:20 AM Marcel Holtmann <marcel@xxxxxxxxxxxx> wrote:
>
> Hi Caleb,
>
> > Since commit 3e54c5890c87 ("Bluetooth: hci_event: Use of a function table to handle HCI events"),
> > some devices see warnings being printed for vendor events, e.g.
> >
> > [   75.806141] Bluetooth: hci0: setting up wcn399x
> > [   75.948311] Bluetooth: hci0: unexpected event 0xff length: 14 > 0
> > [   75.955552] Bluetooth: hci0: QCA Product ID   :0x0000000a
> > [   75.961369] Bluetooth: hci0: QCA SOC Version  :0x40010214
> > [   75.967417] Bluetooth: hci0: QCA ROM Version  :0x00000201
> > [   75.973363] Bluetooth: hci0: QCA Patch Version:0x00000001
> > [   76.000289] Bluetooth: hci0: QCA controller version 0x02140201
> > [   76.006727] Bluetooth: hci0: QCA Downloading qca/crbtfw21.tlv
> > [   76.986850] Bluetooth: hci0: unexpected event 0xff length: 3 > 0
> > [   77.013574] Bluetooth: hci0: QCA Downloading qca/oneplus6/crnv21.bin
> > [   77.024302] Bluetooth: hci0: unexpected event 0xff length: 3 > 0
> > [   77.032681] Bluetooth: hci0: unexpected event 0xff length: 3 > 0
> > [   77.040674] Bluetooth: hci0: unexpected event 0xff length: 3 > 0
> > [   77.049251] Bluetooth: hci0: unexpected event 0xff length: 3 > 0
> > [   77.057997] Bluetooth: hci0: unexpected event 0xff length: 3 > 0
> > [   77.066320] Bluetooth: hci0: unexpected event 0xff length: 3 > 0
> > [   77.075065] Bluetooth: hci0: unexpected event 0xff length: 3 > 0
> > [   77.083073] Bluetooth: hci0: unexpected event 0xff length: 3 > 0
> > [   77.091250] Bluetooth: hci0: unexpected event 0xff length: 3 > 0
> > [   77.099417] Bluetooth: hci0: unexpected event 0xff length: 3 > 0
> > [   77.110166] Bluetooth: hci0: unexpected event 0xff length: 3 > 0
> > [   77.118672] Bluetooth: hci0: unexpected event 0xff length: 3 > 0
> > [   77.127449] Bluetooth: hci0: unexpected event 0xff length: 3 > 0
> > [   77.137190] Bluetooth: hci0: unexpected event 0xff length: 3 > 0
> > [   77.146192] Bluetooth: hci0: unexpected event 0xff length: 3 > 0
> > [   77.154242] Bluetooth: hci0: unexpected event 0xff length: 3 > 0
> > [   77.163183] Bluetooth: hci0: unexpected event 0xff length: 3 > 0
> > [   77.171202] Bluetooth: hci0: unexpected event 0xff length: 3 > 0
> > [   77.179364] Bluetooth: hci0: unexpected event 0xff length: 3 > 0
> > [   77.187259] Bluetooth: hci0: unexpected event 0xff length: 3 > 0
> > [   77.198451] Bluetooth: hci0: QCA setup on UART is completed
> >
> > Avoid printing the event length warning for vendor events, this reverts
> > to the previous behaviour where such warnings weren't printed.
> >
> > Fixes: 3e54c5890c87 ("Bluetooth: hci_event: Use of a function table to handle HCI events")
> > Signed-off-by: Caleb Connolly <caleb.connolly@xxxxxxxxxx>
> > ---
> > Changes since v1:
> > * Don't return early! Vendor events still get parsed despite the
> >   warning. I should have looked a little more closely at that...
> > ---
> > net/bluetooth/hci_event.c | 2 +-
> > 1 file changed, 1 insertion(+), 1 deletion(-)
>
> patch has been applied to bluetooth-stable tree.
>
> Regards
>
> Marcel

I believe a proper fix has already been pushed to bluetooth-next:

https://git.kernel.org/pub/scm/linux/kernel/git/bluetooth/bluetooth-next.git/commit/?id=314d8cd2787418c5ac6b02035c344644f47b292b

HCI_EV_VENDOR shall be assumed to be variable length and that also
uses bt_dev_warn_ratelimited to avoid spamming the logs in case it
still fails.

-- 
Luiz Augusto von Dentz