Re: [PATCH] Revert "Bluetooth: Always enable management interface"

[Gustavo Padovan <gustavo@padovan.org>]

Hi Keith,

* Gustavo Padovan <gustavo@padovan.org> [2012-03-30 03:54:56 -0300]:

> Hi Keith,
> 
> * Keith Packard <keithp@keithp.com> [2012-03-28 19:35:44 -0700]:
> 
> > <#part sign=pgpmime>
> > On Wed, 28 Mar 2012 23:28:39 -0300, Gustavo Padovan <gustavo@padovan.org> wrote:
> > 
> > > A consequence is that you'll need to use newer version of bluez, at least 4.99.
> > > Or call the bluetooth daemon (from a pre 4.99 version) with a parameter to
> > > disable the new interface:
> > 
> > That's not OK -- you're breaking user space with this kernel change. I
> > know I get bashed every time I suggest that we 'fix' the kernel and
> > require new user space X bits...
> 
> Can you try the following patch? It should fix the compatibility problem you had.

Did you had chance to test this patch?

	Gustavo

> 
> commit d21c1177b9cf067809ccee2746633cfea3a8b062
> Author: Gustavo Padovan <gustavo@padovan.org>
> Date:   Thu Mar 29 09:47:53 2012 -0300
> 
>     Bluetooth: Fix userspace compatibility issue with mgmt interface
>     
>     To ensure that old user space versions do not accidentally pick up and
>     try to use the management channel, use a different channel number.
>     
>     Reported-by: Keith Packard <keithp@keithp.com>
>     Acked-by: Johan Hedberg <johan.hedberg@intel.com>
>     Acked-by: Marcel Holtmann <marcel@holtmann.org>
>     Signed-off-by: Gustavo Padovan <gustavo@padovan.org>
> 
> diff --git a/include/net/bluetooth/hci.h b/include/net/bluetooth/hci.h
> index 344b0f9..ba7f148 100644
> --- a/include/net/bluetooth/hci.h
> +++ b/include/net/bluetooth/hci.h
> @@ -1327,8 +1327,8 @@ struct sockaddr_hci {
>  #define HCI_DEV_NONE   0xffff
>  
>  #define HCI_CHANNEL_RAW                0
> -#define HCI_CHANNEL_CONTROL    1
>  #define HCI_CHANNEL_MONITOR    2
> +#define HCI_CHANNEL_CONTROL    3
>  
>  struct hci_filter {
>         unsigned long type_mask;