qemu/audio/audio.c
Hans de Goede b4350deed6 audio: honor QEMU_AUDIO_TIMER_PERIOD instead of waking up every *nano* second
Now that we no longer have MIN_REARM_TIMER_NS a bug in the audio subsys has
clearly shown it self by trying to make a timer fire every nano second.

Note we have a similar problem in 1.6, 1.5 and older but there
MIN_REARM_TIMER_NS limits the wakeups caused by audio being active to
4000 times / second. This still causes a host cpu load of 50 % for simply
playing audio, where as with this patch git master is at 13%, so we should
backport this to 1.5 and 1.6 too.

Note this will not apply to 1.5 and 1.6 as is.

Cc: qemu-stable@nongnu.org
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
2013-10-23 10:37:27 +02:00

2094 lines
51 KiB
C

/*
* QEMU Audio subsystem
*
* Copyright (c) 2003-2005 Vassili Karpov (malc)
*
* Permission is hereby granted, free of charge, to any person obtaining a copy
* of this software and associated documentation files (the "Software"), to deal
* in the Software without restriction, including without limitation the rights
* to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
* copies of the Software, and to permit persons to whom the Software is
* furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
* THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
* THE SOFTWARE.
*/
#include "hw/hw.h"
#include "audio.h"
#include "monitor/monitor.h"
#include "qemu/timer.h"
#include "sysemu/sysemu.h"
#define AUDIO_CAP "audio"
#include "audio_int.h"
/* #define DEBUG_PLIVE */
/* #define DEBUG_LIVE */
/* #define DEBUG_OUT */
/* #define DEBUG_CAPTURE */
/* #define DEBUG_POLL */
#define SW_NAME(sw) (sw)->name ? (sw)->name : "unknown"
/* Order of CONFIG_AUDIO_DRIVERS is import.
The 1st one is the one used by default, that is the reason
that we generate the list.
*/
static struct audio_driver *drvtab[] = {
#ifdef CONFIG_SPICE
&spice_audio_driver,
#endif
CONFIG_AUDIO_DRIVERS
&no_audio_driver,
&wav_audio_driver
};
struct fixed_settings {
int enabled;
int nb_voices;
int greedy;
struct audsettings settings;
};
static struct {
struct fixed_settings fixed_out;
struct fixed_settings fixed_in;
union {
int hertz;
int64_t ticks;
} period;
int plive;
int log_to_monitor;
int try_poll_in;
int try_poll_out;
} conf = {
.fixed_out = { /* DAC fixed settings */
.enabled = 1,
.nb_voices = 1,
.greedy = 1,
.settings = {
.freq = 44100,
.nchannels = 2,
.fmt = AUD_FMT_S16,
.endianness = AUDIO_HOST_ENDIANNESS,
}
},
.fixed_in = { /* ADC fixed settings */
.enabled = 1,
.nb_voices = 1,
.greedy = 1,
.settings = {
.freq = 44100,
.nchannels = 2,
.fmt = AUD_FMT_S16,
.endianness = AUDIO_HOST_ENDIANNESS,
}
},
.period = { .hertz = 250 },
.plive = 0,
.log_to_monitor = 0,
.try_poll_in = 1,
.try_poll_out = 1,
};
static AudioState glob_audio_state;
const struct mixeng_volume nominal_volume = {
.mute = 0,
#ifdef FLOAT_MIXENG
.r = 1.0,
.l = 1.0,
#else
.r = 1ULL << 32,
.l = 1ULL << 32,
#endif
};
#ifdef AUDIO_IS_FLAWLESS_AND_NO_CHECKS_ARE_REQURIED
#error No its not
#else
static void audio_print_options (const char *prefix,
struct audio_option *opt);
int audio_bug (const char *funcname, int cond)
{
if (cond) {
static int shown;
AUD_log (NULL, "A bug was just triggered in %s\n", funcname);
if (!shown) {
struct audio_driver *d;
shown = 1;
AUD_log (NULL, "Save all your work and restart without audio\n");
AUD_log (NULL, "Please send bug report to av1474@comtv.ru\n");
AUD_log (NULL, "I am sorry\n");
d = glob_audio_state.drv;
if (d) {
audio_print_options (d->name, d->options);
}
}
AUD_log (NULL, "Context:\n");
#if defined AUDIO_BREAKPOINT_ON_BUG
# if defined HOST_I386
# if defined __GNUC__
__asm__ ("int3");
# elif defined _MSC_VER
_asm _emit 0xcc;
# else
abort ();
# endif
# else
abort ();
# endif
#endif
}
return cond;
}
#endif
static inline int audio_bits_to_index (int bits)
{
switch (bits) {
case 8:
return 0;
case 16:
return 1;
case 32:
return 2;
default:
audio_bug ("bits_to_index", 1);
AUD_log (NULL, "invalid bits %d\n", bits);
return 0;
}
}
void *audio_calloc (const char *funcname, int nmemb, size_t size)
{
int cond;
size_t len;
len = nmemb * size;
cond = !nmemb || !size;
cond |= nmemb < 0;
cond |= len < size;
if (audio_bug ("audio_calloc", cond)) {
AUD_log (NULL, "%s passed invalid arguments to audio_calloc\n",
funcname);
AUD_log (NULL, "nmemb=%d size=%zu (len=%zu)\n", nmemb, size, len);
return NULL;
}
return g_malloc0 (len);
}
static char *audio_alloc_prefix (const char *s)
{
const char qemu_prefix[] = "QEMU_";
size_t len, i;
char *r, *u;
if (!s) {
return NULL;
}
len = strlen (s);
r = g_malloc (len + sizeof (qemu_prefix));
u = r + sizeof (qemu_prefix) - 1;
pstrcpy (r, len + sizeof (qemu_prefix), qemu_prefix);
pstrcat (r, len + sizeof (qemu_prefix), s);
for (i = 0; i < len; ++i) {
u[i] = qemu_toupper(u[i]);
}
return r;
}
static const char *audio_audfmt_to_string (audfmt_e fmt)
{
switch (fmt) {
case AUD_FMT_U8:
return "U8";
case AUD_FMT_U16:
return "U16";
case AUD_FMT_S8:
return "S8";
case AUD_FMT_S16:
return "S16";
case AUD_FMT_U32:
return "U32";
case AUD_FMT_S32:
return "S32";
}
dolog ("Bogus audfmt %d returning S16\n", fmt);
return "S16";
}
static audfmt_e audio_string_to_audfmt (const char *s, audfmt_e defval,
int *defaultp)
{
if (!strcasecmp (s, "u8")) {
*defaultp = 0;
return AUD_FMT_U8;
}
else if (!strcasecmp (s, "u16")) {
*defaultp = 0;
return AUD_FMT_U16;
}
else if (!strcasecmp (s, "u32")) {
*defaultp = 0;
return AUD_FMT_U32;
}
else if (!strcasecmp (s, "s8")) {
*defaultp = 0;
return AUD_FMT_S8;
}
else if (!strcasecmp (s, "s16")) {
*defaultp = 0;
return AUD_FMT_S16;
}
else if (!strcasecmp (s, "s32")) {
*defaultp = 0;
return AUD_FMT_S32;
}
else {
dolog ("Bogus audio format `%s' using %s\n",
s, audio_audfmt_to_string (defval));
*defaultp = 1;
return defval;
}
}
static audfmt_e audio_get_conf_fmt (const char *envname,
audfmt_e defval,
int *defaultp)
{
const char *var = getenv (envname);
if (!var) {
*defaultp = 1;
return defval;
}
return audio_string_to_audfmt (var, defval, defaultp);
}
static int audio_get_conf_int (const char *key, int defval, int *defaultp)
{
int val;
char *strval;
strval = getenv (key);
if (strval) {
*defaultp = 0;
val = atoi (strval);
return val;
}
else {
*defaultp = 1;
return defval;
}
}
static const char *audio_get_conf_str (const char *key,
const char *defval,
int *defaultp)
{
const char *val = getenv (key);
if (!val) {
*defaultp = 1;
return defval;
}
else {
*defaultp = 0;
return val;
}
}
void AUD_vlog (const char *cap, const char *fmt, va_list ap)
{
if (conf.log_to_monitor) {
if (cap) {
monitor_printf(default_mon, "%s: ", cap);
}
monitor_vprintf(default_mon, fmt, ap);
}
else {
if (cap) {
fprintf (stderr, "%s: ", cap);
}
vfprintf (stderr, fmt, ap);
}
}
void AUD_log (const char *cap, const char *fmt, ...)
{
va_list ap;
va_start (ap, fmt);
AUD_vlog (cap, fmt, ap);
va_end (ap);
}
static void audio_print_options (const char *prefix,
struct audio_option *opt)
{
char *uprefix;
if (!prefix) {
dolog ("No prefix specified\n");
return;
}
if (!opt) {
dolog ("No options\n");
return;
}
uprefix = audio_alloc_prefix (prefix);
for (; opt->name; opt++) {
const char *state = "default";
printf (" %s_%s: ", uprefix, opt->name);
if (opt->overriddenp && *opt->overriddenp) {
state = "current";
}
switch (opt->tag) {
case AUD_OPT_BOOL:
{
int *intp = opt->valp;
printf ("boolean, %s = %d\n", state, *intp ? 1 : 0);
}
break;
case AUD_OPT_INT:
{
int *intp = opt->valp;
printf ("integer, %s = %d\n", state, *intp);
}
break;
case AUD_OPT_FMT:
{
audfmt_e *fmtp = opt->valp;
printf (
"format, %s = %s, (one of: U8 S8 U16 S16 U32 S32)\n",
state,
audio_audfmt_to_string (*fmtp)
);
}
break;
case AUD_OPT_STR:
{
const char **strp = opt->valp;
printf ("string, %s = %s\n",
state,
*strp ? *strp : "(not set)");
}
break;
default:
printf ("???\n");
dolog ("Bad value tag for option %s_%s %d\n",
uprefix, opt->name, opt->tag);
break;
}
printf (" %s\n", opt->descr);
}
g_free (uprefix);
}
static void audio_process_options (const char *prefix,
struct audio_option *opt)
{
char *optname;
const char qemu_prefix[] = "QEMU_";
size_t preflen, optlen;
if (audio_bug (AUDIO_FUNC, !prefix)) {
dolog ("prefix = NULL\n");
return;
}
if (audio_bug (AUDIO_FUNC, !opt)) {
dolog ("opt = NULL\n");
return;
}
preflen = strlen (prefix);
for (; opt->name; opt++) {
size_t len, i;
int def;
if (!opt->valp) {
dolog ("Option value pointer for `%s' is not set\n",
opt->name);
continue;
}
len = strlen (opt->name);
/* len of opt->name + len of prefix + size of qemu_prefix
* (includes trailing zero) + zero + underscore (on behalf of
* sizeof) */
optlen = len + preflen + sizeof (qemu_prefix) + 1;
optname = g_malloc (optlen);
pstrcpy (optname, optlen, qemu_prefix);
/* copy while upper-casing, including trailing zero */
for (i = 0; i <= preflen; ++i) {
optname[i + sizeof (qemu_prefix) - 1] = qemu_toupper(prefix[i]);
}
pstrcat (optname, optlen, "_");
pstrcat (optname, optlen, opt->name);
def = 1;
switch (opt->tag) {
case AUD_OPT_BOOL:
case AUD_OPT_INT:
{
int *intp = opt->valp;
*intp = audio_get_conf_int (optname, *intp, &def);
}
break;
case AUD_OPT_FMT:
{
audfmt_e *fmtp = opt->valp;
*fmtp = audio_get_conf_fmt (optname, *fmtp, &def);
}
break;
case AUD_OPT_STR:
{
const char **strp = opt->valp;
*strp = audio_get_conf_str (optname, *strp, &def);
}
break;
default:
dolog ("Bad value tag for option `%s' - %d\n",
optname, opt->tag);
break;
}
if (!opt->overriddenp) {
opt->overriddenp = &opt->overridden;
}
*opt->overriddenp = !def;
g_free (optname);
}
}
static void audio_print_settings (struct audsettings *as)
{
dolog ("frequency=%d nchannels=%d fmt=", as->freq, as->nchannels);
switch (as->fmt) {
case AUD_FMT_S8:
AUD_log (NULL, "S8");
break;
case AUD_FMT_U8:
AUD_log (NULL, "U8");
break;
case AUD_FMT_S16:
AUD_log (NULL, "S16");
break;
case AUD_FMT_U16:
AUD_log (NULL, "U16");
break;
case AUD_FMT_S32:
AUD_log (NULL, "S32");
break;
case AUD_FMT_U32:
AUD_log (NULL, "U32");
break;
default:
AUD_log (NULL, "invalid(%d)", as->fmt);
break;
}
AUD_log (NULL, " endianness=");
switch (as->endianness) {
case 0:
AUD_log (NULL, "little");
break;
case 1:
AUD_log (NULL, "big");
break;
default:
AUD_log (NULL, "invalid");
break;
}
AUD_log (NULL, "\n");
}
static int audio_validate_settings (struct audsettings *as)
{
int invalid;
invalid = as->nchannels != 1 && as->nchannels != 2;
invalid |= as->endianness != 0 && as->endianness != 1;
switch (as->fmt) {
case AUD_FMT_S8:
case AUD_FMT_U8:
case AUD_FMT_S16:
case AUD_FMT_U16:
case AUD_FMT_S32:
case AUD_FMT_U32:
break;
default:
invalid = 1;
break;
}
invalid |= as->freq <= 0;
return invalid ? -1 : 0;
}
static int audio_pcm_info_eq (struct audio_pcm_info *info, struct audsettings *as)
{
int bits = 8, sign = 0;
switch (as->fmt) {
case AUD_FMT_S8:
sign = 1;
/* fall through */
case AUD_FMT_U8:
break;
case AUD_FMT_S16:
sign = 1;
/* fall through */
case AUD_FMT_U16:
bits = 16;
break;
case AUD_FMT_S32:
sign = 1;
/* fall through */
case AUD_FMT_U32:
bits = 32;
break;
}
return info->freq == as->freq
&& info->nchannels == as->nchannels
&& info->sign == sign
&& info->bits == bits
&& info->swap_endianness == (as->endianness != AUDIO_HOST_ENDIANNESS);
}
void audio_pcm_init_info (struct audio_pcm_info *info, struct audsettings *as)
{
int bits = 8, sign = 0, shift = 0;
switch (as->fmt) {
case AUD_FMT_S8:
sign = 1;
case AUD_FMT_U8:
break;
case AUD_FMT_S16:
sign = 1;
case AUD_FMT_U16:
bits = 16;
shift = 1;
break;
case AUD_FMT_S32:
sign = 1;
case AUD_FMT_U32:
bits = 32;
shift = 2;
break;
}
info->freq = as->freq;
info->bits = bits;
info->sign = sign;
info->nchannels = as->nchannels;
info->shift = (as->nchannels == 2) + shift;
info->align = (1 << info->shift) - 1;
info->bytes_per_second = info->freq << info->shift;
info->swap_endianness = (as->endianness != AUDIO_HOST_ENDIANNESS);
}
void audio_pcm_info_clear_buf (struct audio_pcm_info *info, void *buf, int len)
{
if (!len) {
return;
}
if (info->sign) {
memset (buf, 0x00, len << info->shift);
}
else {
switch (info->bits) {
case 8:
memset (buf, 0x80, len << info->shift);
break;
case 16:
{
int i;
uint16_t *p = buf;
int shift = info->nchannels - 1;
short s = INT16_MAX;
if (info->swap_endianness) {
s = bswap16 (s);
}
for (i = 0; i < len << shift; i++) {
p[i] = s;
}
}
break;
case 32:
{
int i;
uint32_t *p = buf;
int shift = info->nchannels - 1;
int32_t s = INT32_MAX;
if (info->swap_endianness) {
s = bswap32 (s);
}
for (i = 0; i < len << shift; i++) {
p[i] = s;
}
}
break;
default:
AUD_log (NULL, "audio_pcm_info_clear_buf: invalid bits %d\n",
info->bits);
break;
}
}
}
/*
* Capture
*/
static void noop_conv (struct st_sample *dst, const void *src, int samples)
{
(void) src;
(void) dst;
(void) samples;
}
static CaptureVoiceOut *audio_pcm_capture_find_specific (
struct audsettings *as
)
{
CaptureVoiceOut *cap;
AudioState *s = &glob_audio_state;
for (cap = s->cap_head.lh_first; cap; cap = cap->entries.le_next) {
if (audio_pcm_info_eq (&cap->hw.info, as)) {
return cap;
}
}
return NULL;
}
static void audio_notify_capture (CaptureVoiceOut *cap, audcnotification_e cmd)
{
struct capture_callback *cb;
#ifdef DEBUG_CAPTURE
dolog ("notification %d sent\n", cmd);
#endif
for (cb = cap->cb_head.lh_first; cb; cb = cb->entries.le_next) {
cb->ops.notify (cb->opaque, cmd);
}
}
static void audio_capture_maybe_changed (CaptureVoiceOut *cap, int enabled)
{
if (cap->hw.enabled != enabled) {
audcnotification_e cmd;
cap->hw.enabled = enabled;
cmd = enabled ? AUD_CNOTIFY_ENABLE : AUD_CNOTIFY_DISABLE;
audio_notify_capture (cap, cmd);
}
}
static void audio_recalc_and_notify_capture (CaptureVoiceOut *cap)
{
HWVoiceOut *hw = &cap->hw;
SWVoiceOut *sw;
int enabled = 0;
for (sw = hw->sw_head.lh_first; sw; sw = sw->entries.le_next) {
if (sw->active) {
enabled = 1;
break;
}
}
audio_capture_maybe_changed (cap, enabled);
}
static void audio_detach_capture (HWVoiceOut *hw)
{
SWVoiceCap *sc = hw->cap_head.lh_first;
while (sc) {
SWVoiceCap *sc1 = sc->entries.le_next;
SWVoiceOut *sw = &sc->sw;
CaptureVoiceOut *cap = sc->cap;
int was_active = sw->active;
if (sw->rate) {
st_rate_stop (sw->rate);
sw->rate = NULL;
}
QLIST_REMOVE (sw, entries);
QLIST_REMOVE (sc, entries);
g_free (sc);
if (was_active) {
/* We have removed soft voice from the capture:
this might have changed the overall status of the capture
since this might have been the only active voice */
audio_recalc_and_notify_capture (cap);
}
sc = sc1;
}
}
static int audio_attach_capture (HWVoiceOut *hw)
{
AudioState *s = &glob_audio_state;
CaptureVoiceOut *cap;
audio_detach_capture (hw);
for (cap = s->cap_head.lh_first; cap; cap = cap->entries.le_next) {
SWVoiceCap *sc;
SWVoiceOut *sw;
HWVoiceOut *hw_cap = &cap->hw;
sc = audio_calloc (AUDIO_FUNC, 1, sizeof (*sc));
if (!sc) {
dolog ("Could not allocate soft capture voice (%zu bytes)\n",
sizeof (*sc));
return -1;
}
sc->cap = cap;
sw = &sc->sw;
sw->hw = hw_cap;
sw->info = hw->info;
sw->empty = 1;
sw->active = hw->enabled;
sw->conv = noop_conv;
sw->ratio = ((int64_t) hw_cap->info.freq << 32) / sw->info.freq;
sw->vol = nominal_volume;
sw->rate = st_rate_start (sw->info.freq, hw_cap->info.freq);
if (!sw->rate) {
dolog ("Could not start rate conversion for `%s'\n", SW_NAME (sw));
g_free (sw);
return -1;
}
QLIST_INSERT_HEAD (&hw_cap->sw_head, sw, entries);
QLIST_INSERT_HEAD (&hw->cap_head, sc, entries);
#ifdef DEBUG_CAPTURE
sw->name = g_strdup_printf ("for %p %d,%d,%d",
hw, sw->info.freq, sw->info.bits,
sw->info.nchannels);
dolog ("Added %s active = %d\n", sw->name, sw->active);
#endif
if (sw->active) {
audio_capture_maybe_changed (cap, 1);
}
}
return 0;
}
/*
* Hard voice (capture)
*/
static int audio_pcm_hw_find_min_in (HWVoiceIn *hw)
{
SWVoiceIn *sw;
int m = hw->total_samples_captured;
for (sw = hw->sw_head.lh_first; sw; sw = sw->entries.le_next) {
if (sw->active) {
m = audio_MIN (m, sw->total_hw_samples_acquired);
}
}
return m;
}
int audio_pcm_hw_get_live_in (HWVoiceIn *hw)
{
int live = hw->total_samples_captured - audio_pcm_hw_find_min_in (hw);
if (audio_bug (AUDIO_FUNC, live < 0 || live > hw->samples)) {
dolog ("live=%d hw->samples=%d\n", live, hw->samples);
return 0;
}
return live;
}
int audio_pcm_hw_clip_out (HWVoiceOut *hw, void *pcm_buf,
int live, int pending)
{
int left = hw->samples - pending;
int len = audio_MIN (left, live);
int clipped = 0;
while (len) {
struct st_sample *src = hw->mix_buf + hw->rpos;
uint8_t *dst = advance (pcm_buf, hw->rpos << hw->info.shift);
int samples_till_end_of_buf = hw->samples - hw->rpos;
int samples_to_clip = audio_MIN (len, samples_till_end_of_buf);
hw->clip (dst, src, samples_to_clip);
hw->rpos = (hw->rpos + samples_to_clip) % hw->samples;
len -= samples_to_clip;
clipped += samples_to_clip;
}
return clipped;
}
/*
* Soft voice (capture)
*/
static int audio_pcm_sw_get_rpos_in (SWVoiceIn *sw)
{
HWVoiceIn *hw = sw->hw;
int live = hw->total_samples_captured - sw->total_hw_samples_acquired;
int rpos;
if (audio_bug (AUDIO_FUNC, live < 0 || live > hw->samples)) {
dolog ("live=%d hw->samples=%d\n", live, hw->samples);
return 0;
}
rpos = hw->wpos - live;
if (rpos >= 0) {
return rpos;
}
else {
return hw->samples + rpos;
}
}
int audio_pcm_sw_read (SWVoiceIn *sw, void *buf, int size)
{
HWVoiceIn *hw = sw->hw;
int samples, live, ret = 0, swlim, isamp, osamp, rpos, total = 0;
struct st_sample *src, *dst = sw->buf;
rpos = audio_pcm_sw_get_rpos_in (sw) % hw->samples;
live = hw->total_samples_captured - sw->total_hw_samples_acquired;
if (audio_bug (AUDIO_FUNC, live < 0 || live > hw->samples)) {
dolog ("live_in=%d hw->samples=%d\n", live, hw->samples);
return 0;
}
samples = size >> sw->info.shift;
if (!live) {
return 0;
}
swlim = (live * sw->ratio) >> 32;
swlim = audio_MIN (swlim, samples);
while (swlim) {
src = hw->conv_buf + rpos;
isamp = hw->wpos - rpos;
/* XXX: <= ? */
if (isamp <= 0) {
isamp = hw->samples - rpos;
}
if (!isamp) {
break;
}
osamp = swlim;
if (audio_bug (AUDIO_FUNC, osamp < 0)) {
dolog ("osamp=%d\n", osamp);
return 0;
}
st_rate_flow (sw->rate, src, dst, &isamp, &osamp);
swlim -= osamp;
rpos = (rpos + isamp) % hw->samples;
dst += osamp;
ret += osamp;
total += isamp;
}
if (!(hw->ctl_caps & VOICE_VOLUME_CAP)) {
mixeng_volume (sw->buf, ret, &sw->vol);
}
sw->clip (buf, sw->buf, ret);
sw->total_hw_samples_acquired += total;
return ret << sw->info.shift;
}
/*
* Hard voice (playback)
*/
static int audio_pcm_hw_find_min_out (HWVoiceOut *hw, int *nb_livep)
{
SWVoiceOut *sw;
int m = INT_MAX;
int nb_live = 0;
for (sw = hw->sw_head.lh_first; sw; sw = sw->entries.le_next) {
if (sw->active || !sw->empty) {
m = audio_MIN (m, sw->total_hw_samples_mixed);
nb_live += 1;
}
}
*nb_livep = nb_live;
return m;
}
static int audio_pcm_hw_get_live_out (HWVoiceOut *hw, int *nb_live)
{
int smin;
int nb_live1;
smin = audio_pcm_hw_find_min_out (hw, &nb_live1);
if (nb_live) {
*nb_live = nb_live1;
}
if (nb_live1) {
int live = smin;
if (audio_bug (AUDIO_FUNC, live < 0 || live > hw->samples)) {
dolog ("live=%d hw->samples=%d\n", live, hw->samples);
return 0;
}
return live;
}
return 0;
}
/*
* Soft voice (playback)
*/
int audio_pcm_sw_write (SWVoiceOut *sw, void *buf, int size)
{
int hwsamples, samples, isamp, osamp, wpos, live, dead, left, swlim, blck;
int ret = 0, pos = 0, total = 0;
if (!sw) {
return size;
}
hwsamples = sw->hw->samples;
live = sw->total_hw_samples_mixed;
if (audio_bug (AUDIO_FUNC, live < 0 || live > hwsamples)){
dolog ("live=%d hw->samples=%d\n", live, hwsamples);
return 0;
}
if (live == hwsamples) {
#ifdef DEBUG_OUT
dolog ("%s is full %d\n", sw->name, live);
#endif
return 0;
}
wpos = (sw->hw->rpos + live) % hwsamples;
samples = size >> sw->info.shift;
dead = hwsamples - live;
swlim = ((int64_t) dead << 32) / sw->ratio;
swlim = audio_MIN (swlim, samples);
if (swlim) {
sw->conv (sw->buf, buf, swlim);
if (!(sw->hw->ctl_caps & VOICE_VOLUME_CAP)) {
mixeng_volume (sw->buf, swlim, &sw->vol);
}
}
while (swlim) {
dead = hwsamples - live;
left = hwsamples - wpos;
blck = audio_MIN (dead, left);
if (!blck) {
break;
}
isamp = swlim;
osamp = blck;
st_rate_flow_mix (
sw->rate,
sw->buf + pos,
sw->hw->mix_buf + wpos,
&isamp,
&osamp
);
ret += isamp;
swlim -= isamp;
pos += isamp;
live += osamp;
wpos = (wpos + osamp) % hwsamples;
total += osamp;
}
sw->total_hw_samples_mixed += total;
sw->empty = sw->total_hw_samples_mixed == 0;
#ifdef DEBUG_OUT
dolog (
"%s: write size %d ret %d total sw %d\n",
SW_NAME (sw),
size >> sw->info.shift,
ret,
sw->total_hw_samples_mixed
);
#endif
return ret << sw->info.shift;
}
#ifdef DEBUG_AUDIO
static void audio_pcm_print_info (const char *cap, struct audio_pcm_info *info)
{
dolog ("%s: bits %d, sign %d, freq %d, nchan %d\n",
cap, info->bits, info->sign, info->freq, info->nchannels);
}
#endif
#define DAC
#include "audio_template.h"
#undef DAC
#include "audio_template.h"
/*
* Timer
*/
static int audio_is_timer_needed (void)
{
HWVoiceIn *hwi = NULL;
HWVoiceOut *hwo = NULL;
while ((hwo = audio_pcm_hw_find_any_enabled_out (hwo))) {
if (!hwo->poll_mode) return 1;
}
while ((hwi = audio_pcm_hw_find_any_enabled_in (hwi))) {
if (!hwi->poll_mode) return 1;
}
return 0;
}
static void audio_reset_timer (AudioState *s)
{
if (audio_is_timer_needed ()) {
timer_mod (s->ts,
qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL) + conf.period.ticks);
}
else {
timer_del (s->ts);
}
}
static void audio_timer (void *opaque)
{
audio_run ("timer");
audio_reset_timer (opaque);
}
/*
* Public API
*/
int AUD_write (SWVoiceOut *sw, void *buf, int size)
{
int bytes;
if (!sw) {
/* XXX: Consider options */
return size;
}
if (!sw->hw->enabled) {
dolog ("Writing to disabled voice %s\n", SW_NAME (sw));
return 0;
}
bytes = sw->hw->pcm_ops->write (sw, buf, size);
return bytes;
}
int AUD_read (SWVoiceIn *sw, void *buf, int size)
{
int bytes;
if (!sw) {
/* XXX: Consider options */
return size;
}
if (!sw->hw->enabled) {
dolog ("Reading from disabled voice %s\n", SW_NAME (sw));
return 0;
}
bytes = sw->hw->pcm_ops->read (sw, buf, size);
return bytes;
}
int AUD_get_buffer_size_out (SWVoiceOut *sw)
{
return sw->hw->samples << sw->hw->info.shift;
}
void AUD_set_active_out (SWVoiceOut *sw, int on)
{
HWVoiceOut *hw;
if (!sw) {
return;
}
hw = sw->hw;
if (sw->active != on) {
AudioState *s = &glob_audio_state;
SWVoiceOut *temp_sw;
SWVoiceCap *sc;
if (on) {
hw->pending_disable = 0;
if (!hw->enabled) {
hw->enabled = 1;
if (s->vm_running) {
hw->pcm_ops->ctl_out (hw, VOICE_ENABLE, conf.try_poll_out);
audio_reset_timer (s);
}
}
}
else {
if (hw->enabled) {
int nb_active = 0;
for (temp_sw = hw->sw_head.lh_first; temp_sw;
temp_sw = temp_sw->entries.le_next) {
nb_active += temp_sw->active != 0;
}
hw->pending_disable = nb_active == 1;
}
}
for (sc = hw->cap_head.lh_first; sc; sc = sc->entries.le_next) {
sc->sw.active = hw->enabled;
if (hw->enabled) {
audio_capture_maybe_changed (sc->cap, 1);
}
}
sw->active = on;
}
}
void AUD_set_active_in (SWVoiceIn *sw, int on)
{
HWVoiceIn *hw;
if (!sw) {
return;
}
hw = sw->hw;
if (sw->active != on) {
AudioState *s = &glob_audio_state;
SWVoiceIn *temp_sw;
if (on) {
if (!hw->enabled) {
hw->enabled = 1;
if (s->vm_running) {
hw->pcm_ops->ctl_in (hw, VOICE_ENABLE, conf.try_poll_in);
audio_reset_timer (s);
}
}
sw->total_hw_samples_acquired = hw->total_samples_captured;
}
else {
if (hw->enabled) {
int nb_active = 0;
for (temp_sw = hw->sw_head.lh_first; temp_sw;
temp_sw = temp_sw->entries.le_next) {
nb_active += temp_sw->active != 0;
}
if (nb_active == 1) {
hw->enabled = 0;
hw->pcm_ops->ctl_in (hw, VOICE_DISABLE);
}
}
}
sw->active = on;
}
}
static int audio_get_avail (SWVoiceIn *sw)
{
int live;
if (!sw) {
return 0;
}
live = sw->hw->total_samples_captured - sw->total_hw_samples_acquired;
if (audio_bug (AUDIO_FUNC, live < 0 || live > sw->hw->samples)) {
dolog ("live=%d sw->hw->samples=%d\n", live, sw->hw->samples);
return 0;
}
ldebug (
"%s: get_avail live %d ret %" PRId64 "\n",
SW_NAME (sw),
live, (((int64_t) live << 32) / sw->ratio) << sw->info.shift
);
return (((int64_t) live << 32) / sw->ratio) << sw->info.shift;
}
static int audio_get_free (SWVoiceOut *sw)
{
int live, dead;
if (!sw) {
return 0;
}
live = sw->total_hw_samples_mixed;
if (audio_bug (AUDIO_FUNC, live < 0 || live > sw->hw->samples)) {
dolog ("live=%d sw->hw->samples=%d\n", live, sw->hw->samples);
return 0;
}
dead = sw->hw->samples - live;
#ifdef DEBUG_OUT
dolog ("%s: get_free live %d dead %d ret %" PRId64 "\n",
SW_NAME (sw),
live, dead, (((int64_t) dead << 32) / sw->ratio) << sw->info.shift);
#endif
return (((int64_t) dead << 32) / sw->ratio) << sw->info.shift;
}
static void audio_capture_mix_and_clear (HWVoiceOut *hw, int rpos, int samples)
{
int n;
if (hw->enabled) {
SWVoiceCap *sc;
for (sc = hw->cap_head.lh_first; sc; sc = sc->entries.le_next) {
SWVoiceOut *sw = &sc->sw;
int rpos2 = rpos;
n = samples;
while (n) {
int till_end_of_hw = hw->samples - rpos2;
int to_write = audio_MIN (till_end_of_hw, n);
int bytes = to_write << hw->info.shift;
int written;
sw->buf = hw->mix_buf + rpos2;
written = audio_pcm_sw_write (sw, NULL, bytes);
if (written - bytes) {
dolog ("Could not mix %d bytes into a capture "
"buffer, mixed %d\n",
bytes, written);
break;
}
n -= to_write;
rpos2 = (rpos2 + to_write) % hw->samples;
}
}
}
n = audio_MIN (samples, hw->samples - rpos);
mixeng_clear (hw->mix_buf + rpos, n);
mixeng_clear (hw->mix_buf, samples - n);
}
static void audio_run_out (AudioState *s)
{
HWVoiceOut *hw = NULL;
SWVoiceOut *sw;
while ((hw = audio_pcm_hw_find_any_enabled_out (hw))) {
int played;
int live, free, nb_live, cleanup_required, prev_rpos;
live = audio_pcm_hw_get_live_out (hw, &nb_live);
if (!nb_live) {
live = 0;
}
if (audio_bug (AUDIO_FUNC, live < 0 || live > hw->samples)) {
dolog ("live=%d hw->samples=%d\n", live, hw->samples);
continue;
}
if (hw->pending_disable && !nb_live) {
SWVoiceCap *sc;
#ifdef DEBUG_OUT
dolog ("Disabling voice\n");
#endif
hw->enabled = 0;
hw->pending_disable = 0;
hw->pcm_ops->ctl_out (hw, VOICE_DISABLE);
for (sc = hw->cap_head.lh_first; sc; sc = sc->entries.le_next) {
sc->sw.active = 0;
audio_recalc_and_notify_capture (sc->cap);
}
continue;
}
if (!live) {
for (sw = hw->sw_head.lh_first; sw; sw = sw->entries.le_next) {
if (sw->active) {
free = audio_get_free (sw);
if (free > 0) {
sw->callback.fn (sw->callback.opaque, free);
}
}
}
continue;
}
prev_rpos = hw->rpos;
played = hw->pcm_ops->run_out (hw, live);
if (audio_bug (AUDIO_FUNC, hw->rpos >= hw->samples)) {
dolog ("hw->rpos=%d hw->samples=%d played=%d\n",
hw->rpos, hw->samples, played);
hw->rpos = 0;
}
#ifdef DEBUG_OUT
dolog ("played=%d\n", played);
#endif
if (played) {
hw->ts_helper += played;
audio_capture_mix_and_clear (hw, prev_rpos, played);
}
cleanup_required = 0;
for (sw = hw->sw_head.lh_first; sw; sw = sw->entries.le_next) {
if (!sw->active && sw->empty) {
continue;
}
if (audio_bug (AUDIO_FUNC, played > sw->total_hw_samples_mixed)) {
dolog ("played=%d sw->total_hw_samples_mixed=%d\n",
played, sw->total_hw_samples_mixed);
played = sw->total_hw_samples_mixed;
}
sw->total_hw_samples_mixed -= played;
if (!sw->total_hw_samples_mixed) {
sw->empty = 1;
cleanup_required |= !sw->active && !sw->callback.fn;
}
if (sw->active) {
free = audio_get_free (sw);
if (free > 0) {
sw->callback.fn (sw->callback.opaque, free);
}
}
}
if (cleanup_required) {
SWVoiceOut *sw1;
sw = hw->sw_head.lh_first;
while (sw) {
sw1 = sw->entries.le_next;
if (!sw->active && !sw->callback.fn) {
#ifdef DEBUG_PLIVE
dolog ("Finishing with old voice\n");
#endif
audio_close_out (sw);
}
sw = sw1;
}
}
}
}
static void audio_run_in (AudioState *s)
{
HWVoiceIn *hw = NULL;
while ((hw = audio_pcm_hw_find_any_enabled_in (hw))) {
SWVoiceIn *sw;
int captured, min;
captured = hw->pcm_ops->run_in (hw);
min = audio_pcm_hw_find_min_in (hw);
hw->total_samples_captured += captured - min;
hw->ts_helper += captured;
for (sw = hw->sw_head.lh_first; sw; sw = sw->entries.le_next) {
sw->total_hw_samples_acquired -= min;
if (sw->active) {
int avail;
avail = audio_get_avail (sw);
if (avail > 0) {
sw->callback.fn (sw->callback.opaque, avail);
}
}
}
}
}
static void audio_run_capture (AudioState *s)
{
CaptureVoiceOut *cap;
for (cap = s->cap_head.lh_first; cap; cap = cap->entries.le_next) {
int live, rpos, captured;
HWVoiceOut *hw = &cap->hw;
SWVoiceOut *sw;
captured = live = audio_pcm_hw_get_live_out (hw, NULL);
rpos = hw->rpos;
while (live) {
int left = hw->samples - rpos;
int to_capture = audio_MIN (live, left);
struct st_sample *src;
struct capture_callback *cb;
src = hw->mix_buf + rpos;
hw->clip (cap->buf, src, to_capture);
mixeng_clear (src, to_capture);
for (cb = cap->cb_head.lh_first; cb; cb = cb->entries.le_next) {
cb->ops.capture (cb->opaque, cap->buf,
to_capture << hw->info.shift);
}
rpos = (rpos + to_capture) % hw->samples;
live -= to_capture;
}
hw->rpos = rpos;
for (sw = hw->sw_head.lh_first; sw; sw = sw->entries.le_next) {
if (!sw->active && sw->empty) {
continue;
}
if (audio_bug (AUDIO_FUNC, captured > sw->total_hw_samples_mixed)) {
dolog ("captured=%d sw->total_hw_samples_mixed=%d\n",
captured, sw->total_hw_samples_mixed);
captured = sw->total_hw_samples_mixed;
}
sw->total_hw_samples_mixed -= captured;
sw->empty = sw->total_hw_samples_mixed == 0;
}
}
}
void audio_run (const char *msg)
{
AudioState *s = &glob_audio_state;
audio_run_out (s);
audio_run_in (s);
audio_run_capture (s);
#ifdef DEBUG_POLL
{
static double prevtime;
double currtime;
struct timeval tv;
if (gettimeofday (&tv, NULL)) {
perror ("audio_run: gettimeofday");
return;
}
currtime = tv.tv_sec + tv.tv_usec * 1e-6;
dolog ("Elapsed since last %s: %f\n", msg, currtime - prevtime);
prevtime = currtime;
}
#endif
}
static struct audio_option audio_options[] = {
/* DAC */
{
.name = "DAC_FIXED_SETTINGS",
.tag = AUD_OPT_BOOL,
.valp = &conf.fixed_out.enabled,
.descr = "Use fixed settings for host DAC"
},
{
.name = "DAC_FIXED_FREQ",
.tag = AUD_OPT_INT,
.valp = &conf.fixed_out.settings.freq,
.descr = "Frequency for fixed host DAC"
},
{
.name = "DAC_FIXED_FMT",
.tag = AUD_OPT_FMT,
.valp = &conf.fixed_out.settings.fmt,
.descr = "Format for fixed host DAC"
},
{
.name = "DAC_FIXED_CHANNELS",
.tag = AUD_OPT_INT,
.valp = &conf.fixed_out.settings.nchannels,
.descr = "Number of channels for fixed DAC (1 - mono, 2 - stereo)"
},
{
.name = "DAC_VOICES",
.tag = AUD_OPT_INT,
.valp = &conf.fixed_out.nb_voices,
.descr = "Number of voices for DAC"
},
{
.name = "DAC_TRY_POLL",
.tag = AUD_OPT_BOOL,
.valp = &conf.try_poll_out,
.descr = "Attempt using poll mode for DAC"
},
/* ADC */
{
.name = "ADC_FIXED_SETTINGS",
.tag = AUD_OPT_BOOL,
.valp = &conf.fixed_in.enabled,
.descr = "Use fixed settings for host ADC"
},
{
.name = "ADC_FIXED_FREQ",
.tag = AUD_OPT_INT,
.valp = &conf.fixed_in.settings.freq,
.descr = "Frequency for fixed host ADC"
},
{
.name = "ADC_FIXED_FMT",
.tag = AUD_OPT_FMT,
.valp = &conf.fixed_in.settings.fmt,
.descr = "Format for fixed host ADC"
},
{
.name = "ADC_FIXED_CHANNELS",
.tag = AUD_OPT_INT,
.valp = &conf.fixed_in.settings.nchannels,
.descr = "Number of channels for fixed ADC (1 - mono, 2 - stereo)"
},
{
.name = "ADC_VOICES",
.tag = AUD_OPT_INT,
.valp = &conf.fixed_in.nb_voices,
.descr = "Number of voices for ADC"
},
{
.name = "ADC_TRY_POLL",
.tag = AUD_OPT_BOOL,
.valp = &conf.try_poll_in,
.descr = "Attempt using poll mode for ADC"
},
/* Misc */
{
.name = "TIMER_PERIOD",
.tag = AUD_OPT_INT,
.valp = &conf.period.hertz,
.descr = "Timer period in HZ (0 - use lowest possible)"
},
{
.name = "PLIVE",
.tag = AUD_OPT_BOOL,
.valp = &conf.plive,
.descr = "(undocumented)"
},
{
.name = "LOG_TO_MONITOR",
.tag = AUD_OPT_BOOL,
.valp = &conf.log_to_monitor,
.descr = "Print logging messages to monitor instead of stderr"
},
{ /* End of list */ }
};
static void audio_pp_nb_voices (const char *typ, int nb)
{
switch (nb) {
case 0:
printf ("Does not support %s\n", typ);
break;
case 1:
printf ("One %s voice\n", typ);
break;
case INT_MAX:
printf ("Theoretically supports many %s voices\n", typ);
break;
default:
printf ("Theoretically supports up to %d %s voices\n", nb, typ);
break;
}
}
void AUD_help (void)
{
size_t i;
audio_process_options ("AUDIO", audio_options);
for (i = 0; i < ARRAY_SIZE (drvtab); i++) {
struct audio_driver *d = drvtab[i];
if (d->options) {
audio_process_options (d->name, d->options);
}
}
printf ("Audio options:\n");
audio_print_options ("AUDIO", audio_options);
printf ("\n");
printf ("Available drivers:\n");
for (i = 0; i < ARRAY_SIZE (drvtab); i++) {
struct audio_driver *d = drvtab[i];
printf ("Name: %s\n", d->name);
printf ("Description: %s\n", d->descr);
audio_pp_nb_voices ("playback", d->max_voices_out);
audio_pp_nb_voices ("capture", d->max_voices_in);
if (d->options) {
printf ("Options:\n");
audio_print_options (d->name, d->options);
}
else {
printf ("No options\n");
}
printf ("\n");
}
printf (
"Options are settable through environment variables.\n"
"Example:\n"
#ifdef _WIN32
" set QEMU_AUDIO_DRV=wav\n"
" set QEMU_WAV_PATH=c:\\tune.wav\n"
#else
" export QEMU_AUDIO_DRV=wav\n"
" export QEMU_WAV_PATH=$HOME/tune.wav\n"
"(for csh replace export with setenv in the above)\n"
#endif
" qemu ...\n\n"
);
}
static int audio_driver_init (AudioState *s, struct audio_driver *drv)
{
if (drv->options) {
audio_process_options (drv->name, drv->options);
}
s->drv_opaque = drv->init ();
if (s->drv_opaque) {
audio_init_nb_voices_out (drv);
audio_init_nb_voices_in (drv);
s->drv = drv;
return 0;
}
else {
dolog ("Could not init `%s' audio driver\n", drv->name);
return -1;
}
}
static void audio_vm_change_state_handler (void *opaque, int running,
RunState state)
{
AudioState *s = opaque;
HWVoiceOut *hwo = NULL;
HWVoiceIn *hwi = NULL;
int op = running ? VOICE_ENABLE : VOICE_DISABLE;
s->vm_running = running;
while ((hwo = audio_pcm_hw_find_any_enabled_out (hwo))) {
hwo->pcm_ops->ctl_out (hwo, op, conf.try_poll_out);
}
while ((hwi = audio_pcm_hw_find_any_enabled_in (hwi))) {
hwi->pcm_ops->ctl_in (hwi, op, conf.try_poll_in);
}
audio_reset_timer (s);
}
static void audio_atexit (void)
{
AudioState *s = &glob_audio_state;
HWVoiceOut *hwo = NULL;
HWVoiceIn *hwi = NULL;
while ((hwo = audio_pcm_hw_find_any_out (hwo))) {
SWVoiceCap *sc;
if (hwo->enabled) {
hwo->pcm_ops->ctl_out (hwo, VOICE_DISABLE);
}
hwo->pcm_ops->fini_out (hwo);
for (sc = hwo->cap_head.lh_first; sc; sc = sc->entries.le_next) {
CaptureVoiceOut *cap = sc->cap;
struct capture_callback *cb;
for (cb = cap->cb_head.lh_first; cb; cb = cb->entries.le_next) {
cb->ops.destroy (cb->opaque);
}
}
}
while ((hwi = audio_pcm_hw_find_any_in (hwi))) {
if (hwi->enabled) {
hwi->pcm_ops->ctl_in (hwi, VOICE_DISABLE);
}
hwi->pcm_ops->fini_in (hwi);
}
if (s->drv) {
s->drv->fini (s->drv_opaque);
}
}
static const VMStateDescription vmstate_audio = {
.name = "audio",
.version_id = 1,
.minimum_version_id = 1,
.minimum_version_id_old = 1,
.fields = (VMStateField []) {
VMSTATE_END_OF_LIST()
}
};
static void audio_init (void)
{
size_t i;
int done = 0;
const char *drvname;
VMChangeStateEntry *e;
AudioState *s = &glob_audio_state;
if (s->drv) {
return;
}
QLIST_INIT (&s->hw_head_out);
QLIST_INIT (&s->hw_head_in);
QLIST_INIT (&s->cap_head);
atexit (audio_atexit);
s->ts = timer_new_ns(QEMU_CLOCK_VIRTUAL, audio_timer, s);
if (!s->ts) {
hw_error("Could not create audio timer\n");
}
audio_process_options ("AUDIO", audio_options);
s->nb_hw_voices_out = conf.fixed_out.nb_voices;
s->nb_hw_voices_in = conf.fixed_in.nb_voices;
if (s->nb_hw_voices_out <= 0) {
dolog ("Bogus number of playback voices %d, setting to 1\n",
s->nb_hw_voices_out);
s->nb_hw_voices_out = 1;
}
if (s->nb_hw_voices_in <= 0) {
dolog ("Bogus number of capture voices %d, setting to 0\n",
s->nb_hw_voices_in);
s->nb_hw_voices_in = 0;
}
{
int def;
drvname = audio_get_conf_str ("QEMU_AUDIO_DRV", NULL, &def);
}
if (drvname) {
int found = 0;
for (i = 0; i < ARRAY_SIZE (drvtab); i++) {
if (!strcmp (drvname, drvtab[i]->name)) {
done = !audio_driver_init (s, drvtab[i]);
found = 1;
break;
}
}
if (!found) {
dolog ("Unknown audio driver `%s'\n", drvname);
dolog ("Run with -audio-help to list available drivers\n");
}
}
if (!done) {
for (i = 0; !done && i < ARRAY_SIZE (drvtab); i++) {
if (drvtab[i]->can_be_default) {
done = !audio_driver_init (s, drvtab[i]);
}
}
}
if (!done) {
done = !audio_driver_init (s, &no_audio_driver);
if (!done) {
hw_error("Could not initialize audio subsystem\n");
}
else {
dolog ("warning: Using timer based audio emulation\n");
}
}
if (conf.period.hertz <= 0) {
if (conf.period.hertz < 0) {
dolog ("warning: Timer period is negative - %d "
"treating as zero\n",
conf.period.hertz);
}
conf.period.ticks = 1;
} else {
conf.period.ticks =
muldiv64 (1, get_ticks_per_sec (), conf.period.hertz);
}
e = qemu_add_vm_change_state_handler (audio_vm_change_state_handler, s);
if (!e) {
dolog ("warning: Could not register change state handler\n"
"(Audio can continue looping even after stopping the VM)\n");
}
QLIST_INIT (&s->card_head);
vmstate_register (NULL, 0, &vmstate_audio, s);
}
void AUD_register_card (const char *name, QEMUSoundCard *card)
{
audio_init ();
card->name = g_strdup (name);
memset (&card->entries, 0, sizeof (card->entries));
QLIST_INSERT_HEAD (&glob_audio_state.card_head, card, entries);
}
void AUD_remove_card (QEMUSoundCard *card)
{
QLIST_REMOVE (card, entries);
g_free (card->name);
}
CaptureVoiceOut *AUD_add_capture (
struct audsettings *as,
struct audio_capture_ops *ops,
void *cb_opaque
)
{
AudioState *s = &glob_audio_state;
CaptureVoiceOut *cap;
struct capture_callback *cb;
if (audio_validate_settings (as)) {
dolog ("Invalid settings were passed when trying to add capture\n");
audio_print_settings (as);
goto err0;
}
cb = audio_calloc (AUDIO_FUNC, 1, sizeof (*cb));
if (!cb) {
dolog ("Could not allocate capture callback information, size %zu\n",
sizeof (*cb));
goto err0;
}
cb->ops = *ops;
cb->opaque = cb_opaque;
cap = audio_pcm_capture_find_specific (as);
if (cap) {
QLIST_INSERT_HEAD (&cap->cb_head, cb, entries);
return cap;
}
else {
HWVoiceOut *hw;
CaptureVoiceOut *cap;
cap = audio_calloc (AUDIO_FUNC, 1, sizeof (*cap));
if (!cap) {
dolog ("Could not allocate capture voice, size %zu\n",
sizeof (*cap));
goto err1;
}
hw = &cap->hw;
QLIST_INIT (&hw->sw_head);
QLIST_INIT (&cap->cb_head);
/* XXX find a more elegant way */
hw->samples = 4096 * 4;
hw->mix_buf = audio_calloc (AUDIO_FUNC, hw->samples,
sizeof (struct st_sample));
if (!hw->mix_buf) {
dolog ("Could not allocate capture mix buffer (%d samples)\n",
hw->samples);
goto err2;
}
audio_pcm_init_info (&hw->info, as);
cap->buf = audio_calloc (AUDIO_FUNC, hw->samples, 1 << hw->info.shift);
if (!cap->buf) {
dolog ("Could not allocate capture buffer "
"(%d samples, each %d bytes)\n",
hw->samples, 1 << hw->info.shift);
goto err3;
}
hw->clip = mixeng_clip
[hw->info.nchannels == 2]
[hw->info.sign]
[hw->info.swap_endianness]
[audio_bits_to_index (hw->info.bits)];
QLIST_INSERT_HEAD (&s->cap_head, cap, entries);
QLIST_INSERT_HEAD (&cap->cb_head, cb, entries);
hw = NULL;
while ((hw = audio_pcm_hw_find_any_out (hw))) {
audio_attach_capture (hw);
}
return cap;
err3:
g_free (cap->hw.mix_buf);
err2:
g_free (cap);
err1:
g_free (cb);
err0:
return NULL;
}
}
void AUD_del_capture (CaptureVoiceOut *cap, void *cb_opaque)
{
struct capture_callback *cb;
for (cb = cap->cb_head.lh_first; cb; cb = cb->entries.le_next) {
if (cb->opaque == cb_opaque) {
cb->ops.destroy (cb_opaque);
QLIST_REMOVE (cb, entries);
g_free (cb);
if (!cap->cb_head.lh_first) {
SWVoiceOut *sw = cap->hw.sw_head.lh_first, *sw1;
while (sw) {
SWVoiceCap *sc = (SWVoiceCap *) sw;
#ifdef DEBUG_CAPTURE
dolog ("freeing %s\n", sw->name);
#endif
sw1 = sw->entries.le_next;
if (sw->rate) {
st_rate_stop (sw->rate);
sw->rate = NULL;
}
QLIST_REMOVE (sw, entries);
QLIST_REMOVE (sc, entries);
g_free (sc);
sw = sw1;
}
QLIST_REMOVE (cap, entries);
g_free (cap);
}
return;
}
}
}
void AUD_set_volume_out (SWVoiceOut *sw, int mute, uint8_t lvol, uint8_t rvol)
{
if (sw) {
HWVoiceOut *hw = sw->hw;
sw->vol.mute = mute;
sw->vol.l = nominal_volume.l * lvol / 255;
sw->vol.r = nominal_volume.r * rvol / 255;
if (hw->pcm_ops->ctl_out) {
hw->pcm_ops->ctl_out (hw, VOICE_VOLUME, sw);
}
}
}
void AUD_set_volume_in (SWVoiceIn *sw, int mute, uint8_t lvol, uint8_t rvol)
{
if (sw) {
HWVoiceIn *hw = sw->hw;
sw->vol.mute = mute;
sw->vol.l = nominal_volume.l * lvol / 255;
sw->vol.r = nominal_volume.r * rvol / 255;
if (hw->pcm_ops->ctl_in) {
hw->pcm_ops->ctl_in (hw, VOICE_VOLUME, sw);
}
}
}