Re: [Jack-Devel] backend switching - another way

On Tue, Feb 22, 2011 at 10:30:29PM +0000, Peter Nelson wrote:
> On Mon, 2011-02-21 at 22:44 +0000, Fons Adriaensen wrote:
> > On Mon, Feb 21, 2011 at 11:33:26PM +0100, Arnold Krille wrote:
> >  
> > > You seemed to limit your persistent ports to hardware-ports. I like to remove 
> > > the distinction between backends and clients and thus between hardware and 
> > > software ports. And I think that virtual/persitent ports can be benefitial to 
> > > other apps/use-cases as well as for abstracting the actual hardware/backend.
> > 
> > I agree, but as my original post should have made clear, that
> > is not what I proposed to discuss even if I like the idea.
> > Abstracting the HW in this way is something that *could* be
> > done given a bit of courage. Doing this with application ports
> > as well is much more ambitious, it would have deep consequences
> > and would leave little of Jack as we know it.
> >  
> > Ciao,
> > 
> 
> Couldn't you just create those abstract ports with yet another client?
> 
> Perhaps as an internal client (I assume this saves on context
> switches?), it can contain whatever logic for connecting to the real HW
> ports as the user desires, or be controlled by a session manager.
> 
> This might be a reasonable use of jack_port_tie() perhaps...

if this was using a facility like jack_port_tie() it wouldnt need a
process_cb, and we would still get zero copy behaviour.

the impact on the code would be quite small.
(only need to fix jack_port_tie() )

the cons of this approach is that the original h/w ports are still
visible, and autoconnect clients like mplayer and portaudio would still
connect to the system ports, and not the new virtual ones.

nedko already has somthing like this in his ladish rooms
(he is using a separate client, which does memcpy in the process_cb, but
the principle is pretty similar)





> 
> Jack-Devel mailing list
> [hidden]
> http://lists.jackaudio.org/listinfo.cgi/jack-devel-jackaudio.org


-- 
torben Hohn

1298454729.13120_0.ltw:2,a <20110223095152.GA5404 at siel dot b>