[pulseaudio-discuss] Handling JACK driver switching

[Peter Nelson]

On Mon, 2013-01-28 at 07:47 +0200, Tanu Kaskinen wrote:
> Please use the "Reply to All" functionality of your mail client when
> having discussions on a mailing list. I've added pulseaudio-discuss to
> CC again.
> 
> On Mon, 2013-01-28 at 02:36 +0000, Nathan Hulse wrote:
> > The pulse hooks are just the default ones in /usr/lib/pm-utils/sleep.d
> > that ship with pulse 2.0.
> 
> Ok, I wasn't aware of this directory (I have it too, though). If you
> have pulseaudio files there, those are provided by your distribution.
> Upstream pulseaudio doesn't ship any pm scripts.
> 
> > The jack backend is switched from firewire to dummy before suspend, so
> > that I can unload firewire_ohci from the kernel. I load firewire_ohci
> > again at resume and then switch the jack backend to firewire again.
> > But this all happens inbetween pulse saving and restoring state.
> > 
> > 
> > Manually recreating the connection with the jack tools would be a pain
> > in the arse after every suspend/resume. The best solution I have found
> > thus far is to suspend the jack sink/source using pacmd, and then
> > reload module-jackdbus-detect in order to recreate the connections.
> 
> Can't the connection recreation be scripted? But anyway, if reloading
> module-jackdbus-detect works ok for you, that's fine too. Just one
> thing: does unloading module-jackdbus-detect cause the old jack sink and
> source getting unloaded? I'd expect not, so don't you get duplicate jack
> sinks and sources after reloading module-jackdbus-detect?
> 

Hi,

Bit tricky to follow with half the conversation missing, but:

If JACK backend switching causes existing connections to be removed,
then it is the responsibility of the script that changes the backend to
save and restore connections. This problem will apply to all JACK
clients, not just pulseaudio.

Peter.