https://bugs.freedesktop.org/show_bug.cgi?id=87081
--- Comment #6 from David Henningsson <[email protected]> --- (In reply to Tanu Kaskinen from comment #5) > BlueZ 5 doesn't have the Audio interface, which we use with BlueZ 4 for > synchronization (the Audio interface becomes connected after both A2DP and > HSP have become connected, so we wait until the Audio interface become > connected before loading the device module). > > We've had problems with this before, and I think it became clear last time > that it's really not a good idea to implement a synchronization point like > BlueZ 4 has done. Bluetooth devices can and sometimes do advertise their > supported devices asynchronously, in a way that makes it simply impossible > for bluetoothd to know when all profiles have been initialized. > > The problem was solved by extending module-card-restore so that if it tries > to restore a profile that isn't available when the card is initially > created, restoring the profile will be done at the time the profile is > added. Maybe this dynamic profile adding has broken at some point? You mean when the profile is added, or when it becomes available? If we don't have a synchronisation point, then maybe we should just instead not set the profile to "off" unless we really have to. Right now there is code that sets the profile to "off" if the profile has no transport when the card is created, which seems to always be the case to a2dp. What if we instead allow the card to stay at the broken profile? -- You are receiving this mail because: You are the QA Contact for the bug. You are the assignee for the bug.
_______________________________________________ pulseaudio-bugs mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/pulseaudio-bugs
