[alsa-devel] snd_seq_query_port_subscribers not updating subs index

Takashi Iwai tiwai at suse.de
Thu Apr 12 16:54:48 CEST 2018


On Thu, 12 Apr 2018 16:49:27 +0200,
Charles Flèche wrote:
> 
> > Sort of, yes
> 
> >From the snd_seq_query_port_subscribers doc: "To list up all the
> subscriptions, call this function with the index numbers from 0 until this
> returns a negative value."
> 
> So does it mean than in certain cases snd_seq_query_port_subscribers
> increments the index itself, and sometime it does not ?

It shouldn't increment the index number by itself; at least the kernel
interface doesn't do that.  But the API doesn't check whether the
backend (plugin) didn't touch the returned index value or not, so the
possibility is still open :)


Takashi

> 
> Thanks, Takashi
> 
> On Thu, Apr 12, 2018 at 4:44 PM, Takashi Iwai <tiwai at suse.de> wrote:
> 
> > On Sun, 08 Apr 2018 18:37:58 +0200,
> > Charles Flèche wrote:
> > >
> > > Hi all,
> > >
> > > I'm iterating through subscribers with snd_seq_query_port_subscribers to
> > > disconnect them. It works most of the time, but when a2j (an alsa -> jack
> > > MIDI bridge) runs, snd_seq_query_port_subscribers does not update the
> > subs
> > > index, and the next call returns 0, leading to an infinite loop. I have
> > to
> > > manually increase the index when the disconnection fails (meaning that I
> > > hit an a2j connection).
> > >
> > > while (snd_seq_query_port_subscribers(seq, subs) >= 0) {
> > >   addr = *snd_seq_query_subscribe_get_addr(subs);
> > >   if (snd_seq_disconnect_to(seq, port, addr.client, addr.port) < 0) {
> > >     snd_seq_query_subscribe_set_index(subs,
> > > snd_seq_query_subscribe_get_index(subs) + 1);
> > >   }
> > > }
> > >
> > > 1. Is the non-updated subs index an expected behavior ?
> >
> > Sort of, yes.  It's no iterator operation like query_next_client.
> > It just queries the given index and fills out the result.  The user
> > needs to iterate with incrementing the index manually.
> >
> > > 2. Is my fix a correct one or could it leads to other issues ?
> >
> > It's the correct implementation.
> >
> >
> > HTH,
> >
> > Takashi
> >
> [2  <text/html; UTF-8 (quoted-printable)>]
> 


More information about the Alsa-devel mailing list