Re: [PATCH 1/4] audio: remove unnecessary method call

[Johan Hedberg <johan.hedberg@gmail.com>]

Hi,

On Thu, Dec 15, 2011, Luiz Augusto von Dentz wrote:
> On Thu, Dec 15, 2011 at 12:01 PM, Mikel Astiz <mikel.astiz.oss@gmail.com> wrote:
> > From: Mikel Astiz <mikel.astiz@bmw-carit.de>
> >
> > Calling gateway_suspend_stream should not be necessary given that
> > gateway_unlock with rw flags already does it.
> > ---
> >  audio/transport.c |    1 -
> >  1 files changed, 0 insertions(+), 1 deletions(-)
> >
> > diff --git a/audio/transport.c b/audio/transport.c
> > index 7bde32d..6029528 100644
> > --- a/audio/transport.c
> > +++ b/audio/transport.c
> > @@ -535,7 +535,6 @@ static guint suspend_gateway(struct media_transport *transport,
> >                return 0;
> >        }
> >
> > -       gateway_suspend_stream(device);
> >        gateway_unlock(device, GATEWAY_LOCK_READ | GATEWAY_LOCK_WRITE);
> >        g_idle_add(gateway_suspend_complete, owner);
> >        return id++;
> > --
> 
> gateway_suspend_stream as it is useless, but I think we should fix it
> to work similarly to headset_suspend_stream so we properly wait until
> the socket is closed before we unlock and reply, this is necessary in
> order to properly synchronize switching profiles otherwise
> gateway_suspend_complete maybe called without SCO being completely
> disconnected. Also when a fd is passed over to another process calling
> close/g_io_channel_shutdown is not enough since this only release our
> reference, in this case shutdown must be called.

The subsequent patches in the set fail to apply without this first one,
so the whole patch set is on hold until we get an update for it.

Johan