[pulseaudio-discuss] [PATCH v0] bluetooth: Fix potential assertion failure during pairing

[Mikel Astiz]

Hi,

On Mon, Feb 11, 2013 at 11:42 AM, Mikel Astiz <mikel.astiz.oss at gmail.com> wrote:
> From: Mikel Astiz <mikel.astiz at bmw-carit.de>
>
> Commit 40329acc1a28145643e49207e9d65cd05bbda2c8 introduced a potential
> issue during headset pairing. If a headset supporting both HSP/HFP and
> A2DP is paired, and a late UUID (i.e. A2DP) is announced after HSP/HFP
> is already connected, the code in uuid_added_cb() can result in
> duplicated port creation and therefore an assertion failure:
>
> D: [pulseaudio] bluetooth-util.c: dbus: interface=org.bluez.Headset, path=/org/bluez/12840/hci0/dev_00_13_1E_1F_F1_53, member=PropertyChanged
> D: [pulseaudio] bluetooth-util.c: dbus: interface=org.bluez.MediaTransport, path=/org/bluez/12840/hci0/dev_00_13_1E_1F_F1_53/fd0, member=PropertyChanged
> D: [pulseaudio] bluetooth-util.c: dbus: interface=org.bluez.Device, path=/org/bluez/12840/hci0/dev_00_13_1E_1F_F1_53, member=PropertyChanged
> D: [pulseaudio] protocol-dbus.c: Interface org.PulseAudio.Core1.CardProfile added for object /org/pulseaudio/core1/card3/profile2
> E: [pulseaudio] card.c: Assertion 'pa_hashmap_put(c->ports, p->name, p) >= 0' failed at pulsecore/card.c:107, function pa_card_add_ports(). Aborting.
>
> The issue is very unlikely to reproduce but gains interest if further
> ports needs to be merged in the future as well.
> ---
>  src/modules/bluetooth/module-bluetooth-device.c | 42 +++++++++++++++++++++++++
>  1 file changed, 42 insertions(+)
>

Please keep this patch on hold since I'd like to propose another
patchset where all Bluetooth ports would be merged, which would
automatically fix the issue addressed in this patch.

Cheers,
Mikel