[RFC] patch for bluetooth audio.conf

[RFC] patch for bluetooth audio.conf

[Klaus 'mrmoku' Kurzmann]

[-- Attachment #1: Type: text/plain, Size: 883 bytes --]

Hello all,

to get audio routing for bluetooth headsets correct on OpenMoko phones we need 
the attached patch to bluez.inc which changes audio.conf to enables the 
SCORouting=PCM option.

Is this something that can be done globally? Or will it break on other 
hardware? If it cannot be done globally, what is the correct way to do that? 
We (SHR distro) could do that with a distro override, but if it has the 
possibility to break other hardware we would not want that ;) We would rather 
add a machine override then. Right now bluez4 is *not* machine dependent and 
the config files are included in the base package. I see two ways to do that:

a) make bluez4 machine dependent and add overrides for om-gta01/2
b) extract the config files into some bluez4-config package and provide a 
different one for om-gta01/2

Thoughts?
-- 

Klaus 'mrmoku' Kurzmann



[-- Attachment #2: bluez4.patch --]
[-- Type: text/x-patch, Size: 542 bytes --]

diff --git a/recipes/bluez/bluez4.inc b/recipes/bluez/bluez4.inc
index 4c3bfb8..4cb3bc8 100644
--- a/recipes/bluez/bluez4.inc
+++ b/recipes/bluez/bluez4.inc
@@ -41,6 +41,7 @@ EXTRA_OECONF = "\
 "
 
 do_install_append() {
+        sed -i 's/#SCORouting=PCM/SCORouting=PCM/' ${S}/audio/audio.conf
         install -m 0644 ${S}/audio/audio.conf ${D}/${sysconfdir}/bluetooth/
         install -m 0644 ${S}/network/network.conf ${D}/${sysconfdir}/bluetooth/
         install -m 0644 ${S}/input/input.conf ${D}/${sysconfdir}/bluetooth/

Re: [RFC] patch for bluetooth audio.conf

[Phil Blundell]

On Wed, 2009-07-15 at 16:07 +0200, Klaus 'mrmoku' Kurzmann wrote:
> to get audio routing for bluetooth headsets correct on OpenMoko phones we need 
> the attached patch to bluez.inc which changes audio.conf to enables the 
> SCORouting=PCM option.
> 
> Is this something that can be done globally? Or will it break on other 
> hardware? 

It will break on (some) other hardware.  SCORouting is a binary setting:
you can send audio over the HCI, or through a dedicated port (PCM) but
not both.  This change would lose on machines where the PCM interface
isn't connected to anything: in particular that probably means any
machine where the bluetooth hardware is on a usb dongle or CF card.

> If it cannot be done globally, what is the correct way to do that? 
> We (SHR distro) could do that with a distro override, but if it has the 
> possibility to break other hardware we would not want that ;) We would rather 
> add a machine override then. Right now bluez4 is *not* machine dependent and 
> the config files are included in the base package. I see two ways to do that:
> 
> a) make bluez4 machine dependent and add overrides for om-gta01/2
> b) extract the config files into some bluez4-config package and provide a 
> different one for om-gta01/2

Option (a) sounds fairly unattractive.  My first preference would be to
figure this out at runtime (eg by extending blueprobe to do it) rather
than in a static file; if that's impossible or impractical then your
option (b) is probably fine.

p.

Re: [RFC] patch for bluetooth audio.conf

[Klaus 'mrmoku' Kurzmann]

Am Mittwoch 15 Juli 2009 16:16:13 schrieb Phil Blundell:
> On Wed, 2009-07-15 at 16:07 +0200, Klaus 'mrmoku' Kurzmann wrote:
> > to get audio routing for bluetooth headsets correct on OpenMoko phones we
> > need the attached patch to bluez.inc which changes audio.conf to enables
> > the SCORouting=PCM option.
> >
> > Is this something that can be done globally? Or will it break on other
> > hardware?
>
> It will break on (some) other hardware.  SCORouting is a binary setting:
> you can send audio over the HCI, or through a dedicated port (PCM) but
> not both.  This change would lose on machines where the PCM interface
> isn't connected to anything: in particular that probably means any
> machine where the bluetooth hardware is on a usb dongle or CF card.
>
> > If it cannot be done globally, what is the correct way to do that?
> > We (SHR distro) could do that with a distro override, but if it has the
> > possibility to break other hardware we would not want that ;) We would
> > rather add a machine override then. Right now bluez4 is *not* machine
> > dependent and the config files are included in the base package. I see
> > two ways to do that:
> >
> > a) make bluez4 machine dependent and add overrides for om-gta01/2
> > b) extract the config files into some bluez4-config package and provide a
> > different one for om-gta01/2
>
> Option (a) sounds fairly unattractive.  My first preference would be to
> figure this out at runtime (eg by extending blueprobe to do it) rather
> than in a static file; if that's impossible or impractical then your
> option (b) is probably fine.
As I'm told that making bluez autodetect that is difficult... Will take route b) 
then and split out the config files into a bluez4-config package. Would it be ok 
to make that package depend on the machine? First I wanted to do an extra 
package (something like a machine dependent bluez4-config-shr), but maybe it is 
of general interest to be able to have different bluez config for different 
machines...

>
> p.

Thanks

-- 
Klaus 'mrmoku' Kurzmann

Re: [RFC] patch for bluetooth audio.conf

[Koen Kooi]

On 15-07-09 19:24, Klaus 'mrmoku' Kurzmann wrote:
> Am Mittwoch 15 Juli 2009 16:16:13 schrieb Phil Blundell:
>> On Wed, 2009-07-15 at 16:07 +0200, Klaus 'mrmoku' Kurzmann wrote:
>>> to get audio routing for bluetooth headsets correct on OpenMoko phones we
>>> need the attached patch to bluez.inc which changes audio.conf to enables
>>> the SCORouting=PCM option.
>>>
>>> Is this something that can be done globally? Or will it break on other
>>> hardware?
>>
>> It will break on (some) other hardware.  SCORouting is a binary setting:
>> you can send audio over the HCI, or through a dedicated port (PCM) but
>> not both.  This change would lose on machines where the PCM interface
>> isn't connected to anything: in particular that probably means any
>> machine where the bluetooth hardware is on a usb dongle or CF card.
>>
>>> If it cannot be done globally, what is the correct way to do that?
>>> We (SHR distro) could do that with a distro override, but if it has the
>>> possibility to break other hardware we would not want that ;) We would
>>> rather add a machine override then. Right now bluez4 is *not* machine
>>> dependent and the config files are included in the base package. I see
>>> two ways to do that:
>>>
>>> a) make bluez4 machine dependent and add overrides for om-gta01/2
>>> b) extract the config files into some bluez4-config package and provide a
>>> different one for om-gta01/2
>>
>> Option (a) sounds fairly unattractive.  My first preference would be to
>> figure this out at runtime (eg by extending blueprobe to do it) rather
>> than in a static file; if that's impossible or impractical then your
>> option (b) is probably fine.
> As I'm told that making bluez autodetect that is difficult...

Blueprobe can already do autodetection and more, so I don't see what the 
problem is.