[cdi-devel] [PATCH] CDI.audio - without "#define TYNDUR"
Kevin Wolf ·
Matt, can you have a quick look at it, too? If you still haven't got around to learning German, ignore the comments for now - there is noone who actually cares about specifications anyway. ;-) On Mon, Oct 05, 2009 at 08:07:58PM +0200, max at tyndur.org wrote: > From: Max Reitz <max at tyndur.org> > > + A new approach on CDI.audio (aka CDI.sound). Up to now, it just > consists of a single C header describing the interface between > operating system and sound driver (this patch does not contain > a "#define TYNDUR" in "include/cdi.h"). Just a comment on writing patch descriptions for a changed patch: Things like "now without #define TYNDUR" are meaningless in the git history. The summary should describe what the patch is doing, not what previous (not committed) patches did wrong. Usually you would put [PATCH v2] into the subject so that it's clear that this is meant to replace the first version. Things like what you changed in v2 you best describe below the --- line, either directly before or after the diffstat. This way we can read the explanation when reviewing the patch, but it won't be included in the git commit comment. > Signed-off-by: Max Reitz <max at tyndur.org> > --- > include/cdi.h | 1 + > include/cdi/audio.h | 147 +++++++++++++++++++++++++++++++++++++++++++++++++++ > 2 files changed, 148 insertions(+), 0 deletions(-) > create mode 100644 include/cdi/audio.h > > diff --git a/include/cdi.h b/include/cdi.h > index d803751..4f5385c 100644 > --- a/include/cdi.h > +++ b/include/cdi.h > @@ -25,6 +25,7 @@ typedef enum { > CDI_NETWORK = 1, > CDI_STORAGE = 2, > CDI_SCSI = 3, > + CDI_AUDIO = 4, > } cdi_device_type_t; > > struct cdi_driver; > diff --git a/include/cdi/audio.h b/include/cdi/audio.h > new file mode 100644 > index 0000000..bd1fb1a > --- /dev/null > +++ b/include/cdi/audio.h > @@ -0,0 +1,147 @@ > +/* > + * Copyright (c) 2009 Max Reitz > + * > + * This program is free software. It comes without any warranty, to > + * the extent permitted by applicable law. You can redistribute it > + * and/or modify it under the terms of the Do What The Fuck You Want > + * To Public License, Version 2, as published by Sam Hocevar. See > + * http://sam.zoy.org/projects/COPYING.WTFPL for more details. > + */ > + > +/** > + * Treiber fuer Soundkarten > + * \defgroup net > + */ > +/*\@{*/ > + > +#ifndef _CDI_AUDIO_H_ > +#define _CDI_AUDIO_H_ > + > +#include <stddef.h> > + > +#include "cdi.h" > +#include "cdi/list.h" > + > +//Moegliche Operation > +typedef enum { > + //Abspielen > + CDI_AUDIO_PLAY = 0, > + //Aufnehmen > + CDI_AUDIO_RECORD, > + > + CDI_AUDIO_NUM_ACTIONS > +} cdi_audio_action_t; > + > +struct cdi_audio_device { > + struct cdi_device dev; > + > + //Unterstuetzte Modi fuer eine bestimmte Operation. Ist dieser Wert NULL > + //oder die Liste ganz einfach leer, dann wird die Operation nicht > + //unterstuetzt. > + cdi_list_t modes[CDI_AUDIO_NUM_ACTIONS]; > +}; This contains cdi_audio_modes, right? Maybe mention that in the comment. > + > +struct cdi_audio_mode { > + //Samplerate > + int sample_rate; Which unit? > + //Anzahl der Channels (1 = Mono; 2 = Stereo; ...) > + int channels; > + //Bits pro Sample > + int bits_per_sample; > +}; > + > +struct cdi_audio_driver { > + struct cdi_driver drv; > + > + /** > + * Aendert den Modus der Soundkarte fuer eine bestimmte Operation > + * > + * @param device Soundkarte > + * @param action Betreffende Operation > + * @param mode Gewuenschter Modus (muss in device->modes[action] enthalten > + * sein!) > + * > + * @return 0, wenn der neue Modus dem angegebenen entspricht, sonst -1. > + */ > + int (*change_mode)(struct cdi_audio_device* device, > + cdi_audio_action_t action, struct cdi_audio_mode* mode); > + > + /** > + * Liefert den Modus der Soundkarte fuer eine bestimmte Operation > + * > + * @param device Soundkarte > + * @param action Betreffende Operation > + * @param mode Pointer zur Struktur, in der die Informationen zum Modus > + * hinterlegt werden sollen > + * > + * @return 0, wenn die Informationen in *mode gespeichert wurden, sonst -1. > + */ > + int (*get_mode)(struct cdi_audio_device* device, cdi_audio_action_t action, > + struct cdi_audio_mode* mode); > + > + /** > + * Legt die Lautstaerke der angegebenen Operation fest > + * > + * @param device Soundkarte > + * @param action Betreffende Operation > + * @param volume Neue Lautstaerke (zwischen 0 und 255, wobei 255 am > + * lautesten ist; bei -1 wird die Lautstaerke nicht > + * veraendert) > + * > + * @return Die neue Lautstaerke (nuetzlich mit volume == -1) > + */ > + int (*set_volume)(struct cdi_audio_device* device, > + cdi_audio_action_t action, int volume); I think we'll want to control more than a single channel. I'd suggest adding an channel parameter and introducing another function that allows enumerating the channels the device supports. > + > + /** > + * Fuehrt eine Operation aus > + * > + * @param device Soundkarte > + * @param action Auszufuehrende Operation > + * @param buffer Speicher, in dem die Operation durchgefuehrt werden soll > + * @param size Groesse des Speichers > + * @param flags Eventuelle Flags, die die Operation genauer festlegen > + * > + * @return 0 bei Erfolg, -1 im Fehlerfall. > + */ > + int (*act)(struct cdi_audio_device* device, cdi_audio_action_t action, > + void* buffer, size_t size, int flags); You define that -1 is returned in error case, but when is the function supposed to fail? When the device is already playing/recording? How would one detect that new data needs to be sent?
Antworten
- [cdi-devel] [PATCH] CDI.audio - without "#define TYNDUR" · Matthew Iselin