Re: [PATCH] Bluetooth: silence a dmesg error message in hci_request.c

[Luiz Augusto von Dentz]

Hi,

On Thu, Nov 17, 2022 at 1:45 PM Gix, Brian <brian.gix@xxxxxxxxx> wrote:
>
> Hi  Mateusz,
>
> On Thu, 2022-11-17 at 22:27 +0100, Mateusz Jończyk wrote:
> > W dniu 17.11.2022 o 21:34, Gix, Brian pisze:
> > > On Wed, 2022-11-16 at 21:28 +0100, Mateusz Jończyk wrote:
> > > > On kernel 6.1-rcX, I have been getting the following dmesg error
> > > > message
> > > > on every boot, resume from suspend and rfkill unblock of the
> > > > Bluetooth
> > > > device:
> > > >
> > > >         Bluetooth: hci0: HCI_REQ-0xfcf0
> > > >
> > > This has a patch that fixes the usage of the deprecated HCI_REQ
> > > mechanism rather than hiding the fact it is being called, as in
> > > this
> > > case.
> > >
> > > I am still waiting for someone to give me a "Tested-By:" tag to
> > > patch:
> > >
> > > [PATCH 1/1] Bluetooth: Convert MSFT filter HCI cmd to hci_sync
> > >
> > > Which will also stop the dmesg error. If you could try that patch,
> > > and
> > > resend it to the list with a Tested-By tag, it can be applied.
> >
> > Hello,
> >
> > I did not receive this patch, as I was not on the CC list; I was not
> > aware of it. I will test it shortly.

You can find the patch here:

https://patchwork.kernel.org/project/bluetooth/patch/20221102175927.401091-2-brian.gix@xxxxxxxxx/

> >
> > Any guidelines how I should test this functionality? I have a Sony
> > Xperia 10 i4113
> > mobile phone with LineageOS 19.1 / Android 12L, which according to
> > the spec supports
> > Bluetooth 5.0. Quick Google search tells me that I should do things
> > like
> >
> >         hcitool lescan
> >
>
> Whatever you were running that produced the
>
> "Bluetooth: hci0: HCI_REQ-0xfcf0"
>
> error in the dmesg log should be sufficient to determine that the error
> log is no longer happening. The HCI call is necessary on some
> platforms, so the absense of other negative behavior should be
> sufficient to verify that the call is still being made.  The code flow
> itself has not changed, and new coding enforces the HCI command
> sequence, so that it is more deterministric than it was with
> hci_request. The hci_request mechanism was an asyncronous request.
>
> > to discover the phone, then use gatttool to list the services, etc.
> >
> > Greetings,
> >
> > Mateusz
> >
>


-- 
Luiz Augusto von Dentz