Re: [Spice-devel] [PATCH xf86-video-qxl] Revise the XSpice audio processing to avoid the use of pthreads.

2014-10-20 Thread Andrew Eikum
A couple thoughts inline below. On Sat, Oct 18, 2014 at 11:37:13AM -0500, Jeremy White wrote: > The initial implementation used a separate thread to drive the audio > playback channel. But if you have adaptive streaming turned on, > you will eventually get a update_client_playback_latency message

Re: [Spice-devel] [PATCH] Make use of the new playback rate interface, if available.

2014-04-28 Thread Andrew Eikum
ession for some reason, should we do the > >same in the newer codepath? > > I don't think it's a valuable code path; in fact, I fear that it may > have been an overlooked mistake. Looking at git blame, it was > included in the initial xspice audio commit: > 9f18384a

[Spice-devel] [xf86-video-qxl PATCH (resend 2)] Implement sending audio to the client from a directory of FIFO queues

2013-03-15 Thread Andrew Eikum
f (!qxl_map_memory (qxl, scrnIndex)) goto out; diff --git a/src/spiceqxl_audio.c b/src/spiceqxl_audio.c new file mode 100644 index 000..3cd80ff --- /dev/null +++ b/src/spiceqxl_audio.c @@ -0,0 +1,343 @@ +/* + * Copyright 2012 Andrew Eikum for CodeWeavers Inc. + * + * Permis

Re: [Spice-devel] [PATCH] Implement sending audio to the client from a directory of FIFO queues

2013-02-12 Thread Andrew Eikum
dir, playback_fifo_dir, sizeof(qxl->playback_fifo_dir)); +else + qxl->playback_fifo_dir[0] = '\0'; +#endif if (!qxl_map_memory (qxl, scrnIndex)) goto out; diff --git a/src/spiceqxl_audio.c b/src/spiceqxl_audio.c new file mode 100644 index 000..3c

[Spice-devel] [PATCH] Implement sending audio to the client from a directory of FIFO queues

2013-02-12 Thread Andrew Eikum
fo_dir, playback_fifo_dir, sizeof(qxl->playback_fifo_dir)); +else +qxl->playback_fifo_dir[0] = '\0'; +#endif if (!qxl_map_memory (qxl, scrnIndex)) goto out; diff --git a/src/spiceqxl_audio.c b/src/spiceqxl_audio.c new file mode 100644 index 000..2e2b0b3 --

Re: [Spice-devel] [PATCH 1/2] server: Don't release SndChannel twice from worker reference

2012-09-28 Thread Andrew Eikum
On Fri, Sep 28, 2012 at 05:27:03PM +0200, Christophe Fergeau wrote: > On Thu, Sep 13, 2012 at 01:21:07PM -0500, Andrew Eikum wrote: > > > > When we release the SndChannel reference during > > snd_disconnect_channel(), we need to set the pointer to NULL so it > > do

[Spice-devel] [spice.git PATCH 2/2] server: Access the correct SndChannel for a given AudioFrame

2012-09-13 Thread Andrew Eikum
The client of _get_samples() holds a ref to the SndChannel, and we should access that SndChannel when _put_samples() is called, not the one that happens to currently be attached to the Interface. --- (Apologies, both of these are meant for spice.git.) server/snd_worker.c | 16 +++-

[Spice-devel] [PATCH 1/2] server: Don't release SndChannel twice from worker reference

2012-09-13 Thread Andrew Eikum
When we release the SndChannel reference during snd_disconnect_channel(), we need to set the pointer to NULL so it doesn't get released again on client reconnect during snd_set_playback_peer(). This can happen when a reference is held from _playback_get_buffer(). --- server/snd_worker.c | 2 +- 1