M7350/external/bluetooth/bluez/android/hal-audio.c
2024-09-09 08:57:42 +00:00

1899 lines
41 KiB
C

/*
* Copyright (C) 2013 Intel Corporation
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*
*/
#include <errno.h>
#include <pthread.h>
#include <poll.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <sys/socket.h>
#include <sys/un.h>
#include <unistd.h>
#include <arpa/inet.h>
#include <fcntl.h>
#include <hardware/audio.h>
#include <hardware/hardware.h>
#include <sbc/sbc.h>
#include "audio-msg.h"
#include "ipc-common.h"
#include "hal-log.h"
#include "hal-msg.h"
#include "../profiles/audio/a2dp-codecs.h"
#include "../src/shared/util.h"
#define FIXED_A2DP_PLAYBACK_LATENCY_MS 25
#define FIXED_BUFFER_SIZE (20 * 512)
#define MAX_FRAMES_IN_PAYLOAD 15
#define MAX_DELAY 100000 /* 100ms */
#define SBC_QUALITY_MIN_BITPOOL 33
#define SBC_QUALITY_STEP 5
static const uint8_t a2dp_src_uuid[] = {
0x00, 0x00, 0x11, 0x0a, 0x00, 0x00, 0x10, 0x00,
0x80, 0x00, 0x00, 0x80, 0x5f, 0x9b, 0x34, 0xfb };
static int listen_sk = -1;
static int audio_sk = -1;
static pthread_t ipc_th = 0;
static pthread_mutex_t sk_mutex = PTHREAD_MUTEX_INITIALIZER;
#if __BYTE_ORDER == __LITTLE_ENDIAN
struct rtp_header {
unsigned cc:4;
unsigned x:1;
unsigned p:1;
unsigned v:2;
unsigned pt:7;
unsigned m:1;
uint16_t sequence_number;
uint32_t timestamp;
uint32_t ssrc;
uint32_t csrc[0];
} __attribute__ ((packed));
struct rtp_payload {
unsigned frame_count:4;
unsigned rfa0:1;
unsigned is_last_fragment:1;
unsigned is_first_fragment:1;
unsigned is_fragmented:1;
} __attribute__ ((packed));
#elif __BYTE_ORDER == __BIG_ENDIAN
struct rtp_header {
unsigned v:2;
unsigned p:1;
unsigned x:1;
unsigned cc:4;
unsigned m:1;
unsigned pt:7;
uint16_t sequence_number;
uint32_t timestamp;
uint32_t ssrc;
uint32_t csrc[0];
} __attribute__ ((packed));
struct rtp_payload {
unsigned is_fragmented:1;
unsigned is_first_fragment:1;
unsigned is_last_fragment:1;
unsigned rfa0:1;
unsigned frame_count:4;
} __attribute__ ((packed));
#else
#error "Unknown byte order"
#endif
struct media_packet {
struct rtp_header hdr;
struct rtp_payload payload;
uint8_t data[0];
};
struct audio_input_config {
uint32_t rate;
uint32_t channels;
audio_format_t format;
};
struct sbc_data {
a2dp_sbc_t sbc;
sbc_t enc;
uint16_t payload_len;
size_t in_frame_len;
size_t in_buf_size;
size_t out_frame_len;
unsigned frame_duration;
unsigned frames_per_packet;
};
static void timespec_add(struct timespec *base, uint64_t time_us,
struct timespec *res)
{
res->tv_sec = base->tv_sec + time_us / 1000000;
res->tv_nsec = base->tv_nsec + (time_us % 1000000) * 1000;
if (res->tv_nsec >= 1000000000) {
res->tv_sec++;
res->tv_nsec -= 1000000000;
}
}
static void timespec_diff(struct timespec *a, struct timespec *b,
struct timespec *res)
{
res->tv_sec = a->tv_sec - b->tv_sec;
res->tv_nsec = a->tv_nsec - b->tv_nsec;
if (res->tv_nsec < 0) {
res->tv_sec--;
res->tv_nsec += 1000000000; /* 1sec */
}
}
static uint64_t timespec_diff_us(struct timespec *a, struct timespec *b)
{
struct timespec res;
timespec_diff(a, b, &res);
return res.tv_sec * 1000000ll + res.tv_nsec / 1000ll;
}
#if defined(ANDROID)
/*
* Bionic does not have clock_nanosleep() prototype in time.h even though
* it provides its implementation.
*/
extern int clock_nanosleep(clockid_t clock_id, int flags,
const struct timespec *request,
struct timespec *remain);
#endif
static int sbc_get_presets(struct audio_preset *preset, size_t *len);
static int sbc_codec_init(struct audio_preset *preset, uint16_t mtu,
void **codec_data);
static int sbc_cleanup(void *codec_data);
static int sbc_get_config(void *codec_data, struct audio_input_config *config);
static size_t sbc_get_buffer_size(void *codec_data);
static size_t sbc_get_mediapacket_duration(void *codec_data);
static ssize_t sbc_encode_mediapacket(void *codec_data, const uint8_t *buffer,
size_t len, struct media_packet *mp,
size_t mp_data_len, size_t *written);
static bool sbc_update_qos(void *codec_data, uint8_t op);
#define QOS_POLICY_DEFAULT 0x00
#define QOS_POLICY_DECREASE 0x01
struct audio_codec {
uint8_t type;
int (*get_presets) (struct audio_preset *preset, size_t *len);
int (*init) (struct audio_preset *preset, uint16_t mtu,
void **codec_data);
int (*cleanup) (void *codec_data);
int (*get_config) (void *codec_data,
struct audio_input_config *config);
size_t (*get_buffer_size) (void *codec_data);
size_t (*get_mediapacket_duration) (void *codec_data);
ssize_t (*encode_mediapacket) (void *codec_data, const uint8_t *buffer,
size_t len, struct media_packet *mp,
size_t mp_data_len, size_t *written);
bool (*update_qos) (void *codec_data, uint8_t op);
};
static const struct audio_codec audio_codecs[] = {
{
.type = A2DP_CODEC_SBC,
.get_presets = sbc_get_presets,
.init = sbc_codec_init,
.cleanup = sbc_cleanup,
.get_config = sbc_get_config,
.get_buffer_size = sbc_get_buffer_size,
.get_mediapacket_duration = sbc_get_mediapacket_duration,
.encode_mediapacket = sbc_encode_mediapacket,
.update_qos = sbc_update_qos,
}
};
#define NUM_CODECS (sizeof(audio_codecs) / sizeof(audio_codecs[0]))
#define MAX_AUDIO_ENDPOINTS NUM_CODECS
struct audio_endpoint {
uint8_t id;
const struct audio_codec *codec;
void *codec_data;
int fd;
struct media_packet *mp;
size_t mp_data_len;
uint16_t seq;
uint32_t samples;
struct timespec start;
bool resync;
};
static struct audio_endpoint audio_endpoints[MAX_AUDIO_ENDPOINTS];
enum a2dp_state_t {
AUDIO_A2DP_STATE_NONE,
AUDIO_A2DP_STATE_STANDBY,
AUDIO_A2DP_STATE_SUSPENDED,
AUDIO_A2DP_STATE_STARTED
};
struct a2dp_stream_out {
struct audio_stream_out stream;
struct audio_endpoint *ep;
enum a2dp_state_t audio_state;
struct audio_input_config cfg;
uint8_t *downmix_buf;
};
struct a2dp_audio_dev {
struct audio_hw_device dev;
struct a2dp_stream_out *out;
};
static const a2dp_sbc_t sbc_presets[] = {
{
.frequency = SBC_SAMPLING_FREQ_44100 | SBC_SAMPLING_FREQ_48000,
.channel_mode = SBC_CHANNEL_MODE_MONO |
SBC_CHANNEL_MODE_DUAL_CHANNEL |
SBC_CHANNEL_MODE_STEREO |
SBC_CHANNEL_MODE_JOINT_STEREO,
.subbands = SBC_SUBBANDS_4 | SBC_SUBBANDS_8,
.allocation_method = SBC_ALLOCATION_SNR |
SBC_ALLOCATION_LOUDNESS,
.block_length = SBC_BLOCK_LENGTH_4 | SBC_BLOCK_LENGTH_8 |
SBC_BLOCK_LENGTH_12 | SBC_BLOCK_LENGTH_16,
.min_bitpool = MIN_BITPOOL,
.max_bitpool = MAX_BITPOOL
},
{
.frequency = SBC_SAMPLING_FREQ_44100,
.channel_mode = SBC_CHANNEL_MODE_JOINT_STEREO,
.subbands = SBC_SUBBANDS_8,
.allocation_method = SBC_ALLOCATION_LOUDNESS,
.block_length = SBC_BLOCK_LENGTH_16,
.min_bitpool = MIN_BITPOOL,
.max_bitpool = MAX_BITPOOL
},
{
.frequency = SBC_SAMPLING_FREQ_48000,
.channel_mode = SBC_CHANNEL_MODE_JOINT_STEREO,
.subbands = SBC_SUBBANDS_8,
.allocation_method = SBC_ALLOCATION_LOUDNESS,
.block_length = SBC_BLOCK_LENGTH_16,
.min_bitpool = MIN_BITPOOL,
.max_bitpool = MAX_BITPOOL
},
};
static int sbc_get_presets(struct audio_preset *preset, size_t *len)
{
int i;
int count;
size_t new_len = 0;
uint8_t *ptr = (uint8_t *) preset;
size_t preset_size = sizeof(*preset) + sizeof(a2dp_sbc_t);
count = sizeof(sbc_presets) / sizeof(sbc_presets[0]);
for (i = 0; i < count; i++) {
preset = (struct audio_preset *) ptr;
if (new_len + preset_size > *len)
break;
preset->len = sizeof(a2dp_sbc_t);
memcpy(preset->data, &sbc_presets[i], preset->len);
new_len += preset_size;
ptr += preset_size;
}
*len = new_len;
return i;
}
static int sbc_freq2int(uint8_t freq)
{
switch (freq) {
case SBC_SAMPLING_FREQ_16000:
return 16000;
case SBC_SAMPLING_FREQ_32000:
return 32000;
case SBC_SAMPLING_FREQ_44100:
return 44100;
case SBC_SAMPLING_FREQ_48000:
return 48000;
default:
return 0;
}
}
static const char *sbc_mode2str(uint8_t mode)
{
switch (mode) {
case SBC_CHANNEL_MODE_MONO:
return "Mono";
case SBC_CHANNEL_MODE_DUAL_CHANNEL:
return "DualChannel";
case SBC_CHANNEL_MODE_STEREO:
return "Stereo";
case SBC_CHANNEL_MODE_JOINT_STEREO:
return "JointStereo";
default:
return "(unknown)";
}
}
static int sbc_blocks2int(uint8_t blocks)
{
switch (blocks) {
case SBC_BLOCK_LENGTH_4:
return 4;
case SBC_BLOCK_LENGTH_8:
return 8;
case SBC_BLOCK_LENGTH_12:
return 12;
case SBC_BLOCK_LENGTH_16:
return 16;
default:
return 0;
}
}
static int sbc_subbands2int(uint8_t subbands)
{
switch (subbands) {
case SBC_SUBBANDS_4:
return 4;
case SBC_SUBBANDS_8:
return 8;
default:
return 0;
}
}
static const char *sbc_allocation2str(uint8_t allocation)
{
switch (allocation) {
case SBC_ALLOCATION_SNR:
return "SNR";
case SBC_ALLOCATION_LOUDNESS:
return "Loudness";
default:
return "(unknown)";
}
}
static void sbc_init_encoder(struct sbc_data *sbc_data)
{
a2dp_sbc_t *in = &sbc_data->sbc;
sbc_t *out = &sbc_data->enc;
sbc_init_a2dp(out, 0L, in, sizeof(*in));
out->endian = SBC_LE;
out->bitpool = in->max_bitpool;
DBG("frequency=%d channel_mode=%s block_length=%d subbands=%d "
"allocation=%s bitpool=%d-%d",
sbc_freq2int(in->frequency),
sbc_mode2str(in->channel_mode),
sbc_blocks2int(in->block_length),
sbc_subbands2int(in->subbands),
sbc_allocation2str(in->allocation_method),
in->min_bitpool, in->max_bitpool);
}
static void sbc_codec_calculate(struct sbc_data *sbc_data)
{
size_t in_frame_len;
size_t out_frame_len;
size_t num_frames;
in_frame_len = sbc_get_codesize(&sbc_data->enc);
out_frame_len = sbc_get_frame_length(&sbc_data->enc);
num_frames = sbc_data->payload_len / out_frame_len;
sbc_data->in_frame_len = in_frame_len;
sbc_data->in_buf_size = num_frames * in_frame_len;
sbc_data->out_frame_len = out_frame_len;
sbc_data->frame_duration = sbc_get_frame_duration(&sbc_data->enc);
sbc_data->frames_per_packet = num_frames;
DBG("in_frame_len=%zu out_frame_len=%zu frames_per_packet=%zu",
in_frame_len, out_frame_len, num_frames);
}
static int sbc_codec_init(struct audio_preset *preset, uint16_t payload_len,
void **codec_data)
{
struct sbc_data *sbc_data;
if (preset->len != sizeof(a2dp_sbc_t)) {
error("SBC: preset size mismatch");
return AUDIO_STATUS_FAILED;
}
sbc_data = calloc(sizeof(struct sbc_data), 1);
if (!sbc_data)
return AUDIO_STATUS_FAILED;
memcpy(&sbc_data->sbc, preset->data, preset->len);
sbc_init_encoder(sbc_data);
sbc_data->payload_len = payload_len;
sbc_codec_calculate(sbc_data);
*codec_data = sbc_data;
return AUDIO_STATUS_SUCCESS;
}
static int sbc_cleanup(void *codec_data)
{
struct sbc_data *sbc_data = (struct sbc_data *) codec_data;
sbc_finish(&sbc_data->enc);
free(codec_data);
return AUDIO_STATUS_SUCCESS;
}
static int sbc_get_config(void *codec_data, struct audio_input_config *config)
{
struct sbc_data *sbc_data = (struct sbc_data *) codec_data;
switch (sbc_data->sbc.frequency) {
case SBC_SAMPLING_FREQ_16000:
config->rate = 16000;
break;
case SBC_SAMPLING_FREQ_32000:
config->rate = 32000;
break;
case SBC_SAMPLING_FREQ_44100:
config->rate = 44100;
break;
case SBC_SAMPLING_FREQ_48000:
config->rate = 48000;
break;
default:
return AUDIO_STATUS_FAILED;
}
config->channels = sbc_data->sbc.channel_mode == SBC_CHANNEL_MODE_MONO ?
AUDIO_CHANNEL_OUT_MONO :
AUDIO_CHANNEL_OUT_STEREO;
config->format = AUDIO_FORMAT_PCM_16_BIT;
return AUDIO_STATUS_SUCCESS;
}
static size_t sbc_get_buffer_size(void *codec_data)
{
struct sbc_data *sbc_data = (struct sbc_data *) codec_data;
return sbc_data->in_buf_size;
}
static size_t sbc_get_mediapacket_duration(void *codec_data)
{
struct sbc_data *sbc_data = (struct sbc_data *) codec_data;
return sbc_data->frame_duration * sbc_data->frames_per_packet;
}
static ssize_t sbc_encode_mediapacket(void *codec_data, const uint8_t *buffer,
size_t len, struct media_packet *mp,
size_t mp_data_len, size_t *written)
{
struct sbc_data *sbc_data = (struct sbc_data *) codec_data;
size_t consumed = 0;
size_t encoded = 0;
uint8_t frame_count = 0;
while (len - consumed >= sbc_data->in_frame_len &&
mp_data_len - encoded >= sbc_data->out_frame_len &&
frame_count < MAX_FRAMES_IN_PAYLOAD) {
ssize_t read;
ssize_t written = 0;
read = sbc_encode(&sbc_data->enc, buffer + consumed,
sbc_data->in_frame_len, mp->data + encoded,
mp_data_len - encoded, &written);
if (read < 0) {
error("SBC: failed to encode block at frame %d (%zd)",
frame_count, read);
break;
}
frame_count++;
consumed += read;
encoded += written;
}
*written = encoded;
mp->payload.frame_count = frame_count;
return consumed;
}
static bool sbc_update_qos(void *codec_data, uint8_t op)
{
struct sbc_data *sbc_data = (struct sbc_data *) codec_data;
uint8_t curr_bitpool = sbc_data->enc.bitpool;
uint8_t new_bitpool = curr_bitpool;
switch (op) {
case QOS_POLICY_DEFAULT:
new_bitpool = sbc_data->sbc.max_bitpool;
break;
case QOS_POLICY_DECREASE:
if (curr_bitpool > SBC_QUALITY_MIN_BITPOOL) {
new_bitpool = curr_bitpool - SBC_QUALITY_STEP;
if (new_bitpool < SBC_QUALITY_MIN_BITPOOL)
new_bitpool = SBC_QUALITY_MIN_BITPOOL;
}
break;
}
if (new_bitpool == curr_bitpool)
return false;
sbc_data->enc.bitpool = new_bitpool;
sbc_codec_calculate(sbc_data);
info("SBC: bitpool changed: %d -> %d", curr_bitpool, new_bitpool);
return true;
}
static int audio_ipc_cmd(uint8_t service_id, uint8_t opcode, uint16_t len,
void *param, size_t *rsp_len, void *rsp, int *fd)
{
ssize_t ret;
struct msghdr msg;
struct iovec iv[2];
struct ipc_hdr cmd;
char cmsgbuf[CMSG_SPACE(sizeof(int))];
struct ipc_status s;
size_t s_len = sizeof(s);
pthread_mutex_lock(&sk_mutex);
if (audio_sk < 0) {
error("audio: Invalid cmd socket passed to audio_ipc_cmd");
goto failed;
}
if (!rsp || !rsp_len) {
memset(&s, 0, s_len);
rsp_len = &s_len;
rsp = &s;
}
memset(&msg, 0, sizeof(msg));
memset(&cmd, 0, sizeof(cmd));
cmd.service_id = service_id;
cmd.opcode = opcode;
cmd.len = len;
iv[0].iov_base = &cmd;
iv[0].iov_len = sizeof(cmd);
iv[1].iov_base = param;
iv[1].iov_len = len;
msg.msg_iov = iv;
msg.msg_iovlen = 2;
ret = sendmsg(audio_sk, &msg, 0);
if (ret < 0) {
error("audio: Sending command failed:%s", strerror(errno));
goto failed;
}
/* socket was shutdown */
if (ret == 0) {
error("audio: Command socket closed");
goto failed;
}
memset(&msg, 0, sizeof(msg));
memset(&cmd, 0, sizeof(cmd));
iv[0].iov_base = &cmd;
iv[0].iov_len = sizeof(cmd);
iv[1].iov_base = rsp;
iv[1].iov_len = *rsp_len;
msg.msg_iov = iv;
msg.msg_iovlen = 2;
if (fd) {
memset(cmsgbuf, 0, sizeof(cmsgbuf));
msg.msg_control = cmsgbuf;
msg.msg_controllen = sizeof(cmsgbuf);
}
ret = recvmsg(audio_sk, &msg, 0);
if (ret < 0) {
error("audio: Receiving command response failed:%s",
strerror(errno));
goto failed;
}
if (ret < (ssize_t) sizeof(cmd)) {
error("audio: Too small response received(%zd bytes)", ret);
goto failed;
}
if (cmd.service_id != service_id) {
error("audio: Invalid service id (%u vs %u)", cmd.service_id,
service_id);
goto failed;
}
if (ret != (ssize_t) (sizeof(cmd) + cmd.len)) {
error("audio: Malformed response received(%zd bytes)", ret);
goto failed;
}
if (cmd.opcode != opcode && cmd.opcode != AUDIO_OP_STATUS) {
error("audio: Invalid opcode received (%u vs %u)",
cmd.opcode, opcode);
goto failed;
}
if (cmd.opcode == AUDIO_OP_STATUS) {
struct ipc_status *s = rsp;
if (sizeof(*s) != cmd.len) {
error("audio: Invalid status length");
goto failed;
}
if (s->code == AUDIO_STATUS_SUCCESS) {
error("audio: Invalid success status response");
goto failed;
}
pthread_mutex_unlock(&sk_mutex);
return s->code;
}
pthread_mutex_unlock(&sk_mutex);
/* Receive auxiliary data in msg */
if (fd) {
struct cmsghdr *cmsg;
*fd = -1;
for (cmsg = CMSG_FIRSTHDR(&msg); cmsg;
cmsg = CMSG_NXTHDR(&msg, cmsg)) {
if (cmsg->cmsg_level == SOL_SOCKET
&& cmsg->cmsg_type == SCM_RIGHTS) {
memcpy(fd, CMSG_DATA(cmsg), sizeof(int));
break;
}
}
if (*fd < 0)
goto failed;
}
if (rsp_len)
*rsp_len = cmd.len;
return AUDIO_STATUS_SUCCESS;
failed:
/* Some serious issue happen on IPC - recover */
shutdown(audio_sk, SHUT_RDWR);
pthread_mutex_unlock(&sk_mutex);
return AUDIO_STATUS_FAILED;
}
static int ipc_open_cmd(const struct audio_codec *codec)
{
uint8_t buf[BLUEZ_AUDIO_MTU];
struct audio_cmd_open *cmd = (struct audio_cmd_open *) buf;
struct audio_rsp_open rsp;
size_t cmd_len = sizeof(buf) - sizeof(*cmd);
size_t rsp_len = sizeof(rsp);
int result;
DBG("");
memcpy(cmd->uuid, a2dp_src_uuid, sizeof(a2dp_src_uuid));
cmd->codec = codec->type;
cmd->presets = codec->get_presets(cmd->preset, &cmd_len);
cmd_len += sizeof(*cmd);
result = audio_ipc_cmd(AUDIO_SERVICE_ID, AUDIO_OP_OPEN, cmd_len, cmd,
&rsp_len, &rsp, NULL);
if (result != AUDIO_STATUS_SUCCESS)
return 0;
return rsp.id;
}
static int ipc_close_cmd(uint8_t endpoint_id)
{
struct audio_cmd_close cmd;
int result;
DBG("");
cmd.id = endpoint_id;
result = audio_ipc_cmd(AUDIO_SERVICE_ID, AUDIO_OP_CLOSE,
sizeof(cmd), &cmd, NULL, NULL, NULL);
return result;
}
static int ipc_open_stream_cmd(uint8_t endpoint_id, uint16_t *mtu, int *fd,
struct audio_preset **caps)
{
char buf[BLUEZ_AUDIO_MTU];
struct audio_cmd_open_stream cmd;
struct audio_rsp_open_stream *rsp =
(struct audio_rsp_open_stream *) &buf;
size_t rsp_len = sizeof(buf);
int result;
DBG("");
if (!caps)
return AUDIO_STATUS_FAILED;
cmd.id = endpoint_id;
result = audio_ipc_cmd(AUDIO_SERVICE_ID, AUDIO_OP_OPEN_STREAM,
sizeof(cmd), &cmd, &rsp_len, rsp, fd);
if (result == AUDIO_STATUS_SUCCESS) {
size_t buf_len = sizeof(struct audio_preset) +
rsp->preset[0].len;
*mtu = rsp->mtu;
*caps = malloc(buf_len);
memcpy(*caps, &rsp->preset, buf_len);
} else {
*caps = NULL;
}
return result;
}
static int ipc_close_stream_cmd(uint8_t endpoint_id)
{
struct audio_cmd_close_stream cmd;
int result;
DBG("");
cmd.id = endpoint_id;
result = audio_ipc_cmd(AUDIO_SERVICE_ID, AUDIO_OP_CLOSE_STREAM,
sizeof(cmd), &cmd, NULL, NULL, NULL);
return result;
}
static int ipc_resume_stream_cmd(uint8_t endpoint_id)
{
struct audio_cmd_resume_stream cmd;
int result;
DBG("");
cmd.id = endpoint_id;
result = audio_ipc_cmd(AUDIO_SERVICE_ID, AUDIO_OP_RESUME_STREAM,
sizeof(cmd), &cmd, NULL, NULL, NULL);
return result;
}
static int ipc_suspend_stream_cmd(uint8_t endpoint_id)
{
struct audio_cmd_suspend_stream cmd;
int result;
DBG("");
cmd.id = endpoint_id;
result = audio_ipc_cmd(AUDIO_SERVICE_ID, AUDIO_OP_SUSPEND_STREAM,
sizeof(cmd), &cmd, NULL, NULL, NULL);
return result;
}
static int register_endpoints(void)
{
struct audio_endpoint *ep = &audio_endpoints[0];
size_t i;
for (i = 0; i < NUM_CODECS; i++, ep++) {
const struct audio_codec *codec = &audio_codecs[i];
ep->id = ipc_open_cmd(codec);
if (!ep->id)
return AUDIO_STATUS_FAILED;
ep->codec = codec;
ep->codec_data = NULL;
ep->fd = -1;
}
return AUDIO_STATUS_SUCCESS;
}
static void unregister_endpoints(void)
{
size_t i;
for (i = 0; i < MAX_AUDIO_ENDPOINTS; i++) {
struct audio_endpoint *ep = &audio_endpoints[i];
if (ep->id) {
ipc_close_cmd(ep->id);
memset(ep, 0, sizeof(*ep));
}
}
}
static bool open_endpoint(struct audio_endpoint *ep,
struct audio_input_config *cfg)
{
struct audio_preset *preset;
const struct audio_codec *codec;
uint16_t mtu;
uint16_t payload_len;
int fd;
if (ipc_open_stream_cmd(ep->id, &mtu, &fd, &preset) !=
AUDIO_STATUS_SUCCESS)
return false;
DBG("mtu=%u", mtu);
payload_len = mtu - sizeof(*ep->mp);
ep->fd = fd;
codec = ep->codec;
codec->init(preset, payload_len, &ep->codec_data);
codec->get_config(ep->codec_data, cfg);
ep->mp = calloc(mtu, 1);
if (!ep->mp)
goto failed;
ep->mp->hdr.v = 2;
ep->mp->hdr.pt = 1;
ep->mp->hdr.ssrc = htonl(1);
ep->mp_data_len = payload_len;
free(preset);
return true;
failed:
close(fd);
free(preset);
return false;
}
static void close_endpoint(struct audio_endpoint *ep)
{
ipc_close_stream_cmd(ep->id);
if (ep->fd >= 0) {
close(ep->fd);
ep->fd = -1;
}
free(ep->mp);
ep->codec->cleanup(ep->codec_data);
ep->codec_data = NULL;
}
static bool resume_endpoint(struct audio_endpoint *ep)
{
if (ipc_resume_stream_cmd(ep->id) != AUDIO_STATUS_SUCCESS)
return false;
ep->samples = 0;
ep->resync = false;
ep->codec->update_qos(ep->codec_data, QOS_POLICY_DEFAULT);
return true;
}
static void downmix_to_mono(struct a2dp_stream_out *out, const uint8_t *buffer,
size_t bytes)
{
const int16_t *input = (const void *) buffer;
int16_t *output = (void *) out->downmix_buf;
size_t i;
for (i = 0; i < bytes / 2; i++) {
int16_t l = le16_to_cpu(get_unaligned(&input[i * 2]));
int16_t r = le16_to_cpu(get_unaligned(&input[i * 2 + 1]));
put_unaligned(cpu_to_le16((l + r) / 2), &output[i]);
}
}
static bool wait_for_endpoint(struct audio_endpoint *ep, bool *writable)
{
int ret;
while (true) {
struct pollfd pollfd;
pollfd.fd = ep->fd;
pollfd.events = POLLOUT;
pollfd.revents = 0;
ret = poll(&pollfd, 1, 500);
if (ret >= 0) {
*writable = !!(pollfd.revents & POLLOUT);
break;
}
if (errno != EINTR) {
ret = errno;
error("poll failed (%d)", ret);
return false;
}
}
return true;
}
static bool write_to_endpoint(struct audio_endpoint *ep, size_t bytes)
{
struct media_packet *mp = (struct media_packet *) ep->mp;
int ret;
while (true) {
ret = write(ep->fd, mp, sizeof(*mp) + bytes);
if (ret >= 0)
break;
/*
* this should not happen so let's issue warning, but do not
* fail, we can try to write next packet
*/
if (errno == EAGAIN) {
ret = errno;
warn("write failed (%d)", ret);
break;
}
if (errno != EINTR) {
ret = errno;
error("write failed (%d)", ret);
return false;
}
}
return true;
}
static bool write_data(struct a2dp_stream_out *out, const void *buffer,
size_t bytes)
{
struct audio_endpoint *ep = out->ep;
struct media_packet *mp = (struct media_packet *) ep->mp;
size_t free_space = ep->mp_data_len;
size_t consumed = 0;
while (consumed < bytes) {
size_t written = 0;
ssize_t read;
uint32_t samples;
int ret;
struct timespec current;
uint64_t audio_sent, audio_passed;
bool do_write = false;
/*
* prepare media packet in advance so we don't waste time after
* wakeup
*/
mp->hdr.sequence_number = htons(ep->seq++);
mp->hdr.timestamp = htonl(ep->samples);
read = ep->codec->encode_mediapacket(ep->codec_data,
buffer + consumed,
bytes - consumed, mp,
free_space, &written);
/*
* not much we can do here, let's just ignore remaining
* data and continue
*/
if (read <= 0)
return true;
/* calculate where are we and where we should be */
clock_gettime(CLOCK_MONOTONIC, &current);
if (!ep->samples)
memcpy(&ep->start, &current, sizeof(ep->start));
audio_sent = ep->samples * 1000000ll / out->cfg.rate;
audio_passed = timespec_diff_us(&current, &ep->start);
/*
* if we're ahead of stream then wait for next write point,
* if we're lagging more than 100ms then stop writing and just
* skip data until we're back in sync
*/
if (audio_sent > audio_passed) {
struct timespec anchor;
ep->resync = false;
timespec_add(&ep->start, audio_sent, &anchor);
while (true) {
ret = clock_nanosleep(CLOCK_MONOTONIC,
TIMER_ABSTIME, &anchor,
NULL);
if (!ret)
break;
if (ret != EINTR) {
error("clock_nanosleep failed (%d)",
ret);
return false;
}
}
} else if (!ep->resync) {
uint64_t diff = audio_passed - audio_sent;
if (diff > MAX_DELAY) {
warn("lag is %jums, resyncing", diff / 1000);
ep->codec->update_qos(ep->codec_data,
QOS_POLICY_DECREASE);
ep->resync = true;
}
}
/* in resync mode we'll just drop mediapackets */
if (!ep->resync) {
/* wait some time for socket to be ready for write,
* but we'll just skip writing data if timeout occurs
*/
if (!wait_for_endpoint(ep, &do_write))
return false;
if (do_write)
if (!write_to_endpoint(ep, written))
return false;
}
/*
* AudioFlinger provides 16bit PCM, so sample size is 2 bytes
* multiplied by number of channels. Number of channels is
* simply number of bits set in channels mask.
*/
samples = read / (2 * popcount(out->cfg.channels));
ep->samples += samples;
consumed += read;
}
return true;
}
static ssize_t out_write(struct audio_stream_out *stream, const void *buffer,
size_t bytes)
{
struct a2dp_stream_out *out = (struct a2dp_stream_out *) stream;
const void *in_buf = buffer;
size_t in_len = bytes;
/* just return in case we're closing */
if (out->audio_state == AUDIO_A2DP_STATE_NONE)
return -1;
/* We can auto-start only from standby */
if (out->audio_state == AUDIO_A2DP_STATE_STANDBY) {
DBG("stream in standby, auto-start");
if (!resume_endpoint(out->ep))
return -1;
out->audio_state = AUDIO_A2DP_STATE_STARTED;
}
if (out->audio_state != AUDIO_A2DP_STATE_STARTED) {
error("audio: stream not started");
return -1;
}
if (out->ep->fd < 0) {
error("audio: no transport socket");
return -1;
}
/*
* currently Android audioflinger is not able to provide mono stream on
* A2DP output so down mixing needs to be done in hal-audio plugin.
*
* for reference see
* AudioFlinger::PlaybackThread::readOutputParameters()
* frameworks/av/services/audioflinger/Threads.cpp:1631
*/
if (out->cfg.channels == AUDIO_CHANNEL_OUT_MONO) {
if (!out->downmix_buf) {
error("audio: downmix buffer not initialized");
return -1;
}
downmix_to_mono(out, buffer, bytes);
in_buf = out->downmix_buf;
in_len = bytes / 2;
}
if (!write_data(out, in_buf, in_len))
return -1;
return bytes;
}
static uint32_t out_get_sample_rate(const struct audio_stream *stream)
{
struct a2dp_stream_out *out = (struct a2dp_stream_out *) stream;
DBG("");
return out->cfg.rate;
}
static int out_set_sample_rate(struct audio_stream *stream, uint32_t rate)
{
struct a2dp_stream_out *out = (struct a2dp_stream_out *) stream;
DBG("");
if (rate != out->cfg.rate) {
warn("audio: cannot set sample rate to %d", rate);
return -1;
}
return 0;
}
static size_t out_get_buffer_size(const struct audio_stream *stream)
{
DBG("");
/*
* We should return proper buffer size calculated by codec (so each
* input buffer is encoded into single media packed) but this does not
* work well with AudioFlinger and causes problems. For this reason we
* use magic value here and out_write code takes care of splitting
* input buffer into multiple media packets.
*/
return FIXED_BUFFER_SIZE;
}
static uint32_t out_get_channels(const struct audio_stream *stream)
{
DBG("");
/*
* AudioFlinger can only provide stereo stream, so we return it here and
* later we'll downmix this to mono in case codec requires it
*/
return AUDIO_CHANNEL_OUT_STEREO;
}
static audio_format_t out_get_format(const struct audio_stream *stream)
{
struct a2dp_stream_out *out = (struct a2dp_stream_out *) stream;
DBG("");
return out->cfg.format;
}
static int out_set_format(struct audio_stream *stream, audio_format_t format)
{
DBG("");
return -ENOSYS;
}
static int out_standby(struct audio_stream *stream)
{
struct a2dp_stream_out *out = (struct a2dp_stream_out *) stream;
DBG("");
if (out->audio_state == AUDIO_A2DP_STATE_STARTED) {
if (ipc_suspend_stream_cmd(out->ep->id) != AUDIO_STATUS_SUCCESS)
return -1;
out->audio_state = AUDIO_A2DP_STATE_STANDBY;
}
return 0;
}
static int out_dump(const struct audio_stream *stream, int fd)
{
DBG("");
return -ENOSYS;
}
static int out_set_parameters(struct audio_stream *stream, const char *kvpairs)
{
struct a2dp_stream_out *out = (struct a2dp_stream_out *) stream;
char *kvpair;
char *str;
char *saveptr;
bool enter_suspend = false;
bool exit_suspend = false;
DBG("%s", kvpairs);
str = strdup(kvpairs);
if (!str)
return -ENOMEM;
kvpair = strtok_r(str, ";", &saveptr);
for (; kvpair && *kvpair; kvpair = strtok_r(NULL, ";", &saveptr)) {
char *keyval;
keyval = strchr(kvpair, '=');
if (!keyval)
continue;
*keyval = '\0';
keyval++;
if (!strcmp(kvpair, "closing")) {
if (!strcmp(keyval, "true"))
out->audio_state = AUDIO_A2DP_STATE_NONE;
} else if (!strcmp(kvpair, "A2dpSuspended")) {
if (!strcmp(keyval, "true"))
enter_suspend = true;
else
exit_suspend = true;
}
}
free(str);
if (enter_suspend && out->audio_state == AUDIO_A2DP_STATE_STARTED) {
if (ipc_suspend_stream_cmd(out->ep->id) != AUDIO_STATUS_SUCCESS)
return -1;
out->audio_state = AUDIO_A2DP_STATE_SUSPENDED;
}
if (exit_suspend && out->audio_state == AUDIO_A2DP_STATE_SUSPENDED)
out->audio_state = AUDIO_A2DP_STATE_STANDBY;
return 0;
}
static char *out_get_parameters(const struct audio_stream *stream,
const char *keys)
{
DBG("");
return strdup("");
}
static uint32_t out_get_latency(const struct audio_stream_out *stream)
{
struct a2dp_stream_out *out = (struct a2dp_stream_out *) stream;
struct audio_endpoint *ep = out->ep;
size_t pkt_duration;
DBG("");
pkt_duration = ep->codec->get_mediapacket_duration(ep->codec_data);
return FIXED_A2DP_PLAYBACK_LATENCY_MS + pkt_duration / 1000;
}
static int out_set_volume(struct audio_stream_out *stream, float left,
float right)
{
DBG("");
/* volume controlled in audioflinger mixer (digital) */
return -ENOSYS;
}
static int out_get_render_position(const struct audio_stream_out *stream,
uint32_t *dsp_frames)
{
DBG("");
return -ENOSYS;
}
static int out_add_audio_effect(const struct audio_stream *stream,
effect_handle_t effect)
{
DBG("");
return -ENOSYS;
}
static int out_remove_audio_effect(const struct audio_stream *stream,
effect_handle_t effect)
{
DBG("");
return -ENOSYS;
}
static uint32_t in_get_sample_rate(const struct audio_stream *stream)
{
DBG("");
return -ENOSYS;
}
static int in_set_sample_rate(struct audio_stream *stream, uint32_t rate)
{
DBG("");
return -ENOSYS;
}
static size_t in_get_buffer_size(const struct audio_stream *stream)
{
DBG("");
return -ENOSYS;
}
static uint32_t in_get_channels(const struct audio_stream *stream)
{
DBG("");
return -ENOSYS;
}
static audio_format_t in_get_format(const struct audio_stream *stream)
{
DBG("");
return -ENOSYS;
}
static int in_set_format(struct audio_stream *stream, audio_format_t format)
{
DBG("");
return -ENOSYS;
}
static int in_standby(struct audio_stream *stream)
{
DBG("");
return -ENOSYS;
}
static int in_dump(const struct audio_stream *stream, int fd)
{
DBG("");
return -ENOSYS;
}
static int in_set_parameters(struct audio_stream *stream, const char *kvpairs)
{
DBG("");
return -ENOSYS;
}
static char *in_get_parameters(const struct audio_stream *stream,
const char *keys)
{
DBG("");
return strdup("");
}
static int in_set_gain(struct audio_stream_in *stream, float gain)
{
DBG("");
return -ENOSYS;
}
static ssize_t in_read(struct audio_stream_in *stream, void *buffer,
size_t bytes)
{
DBG("");
return -ENOSYS;
}
static uint32_t in_get_input_frames_lost(struct audio_stream_in *stream)
{
DBG("");
return -ENOSYS;
}
static int in_add_audio_effect(const struct audio_stream *stream,
effect_handle_t effect)
{
DBG("");
return -ENOSYS;
}
static int in_remove_audio_effect(const struct audio_stream *stream,
effect_handle_t effect)
{
DBG("");
return -ENOSYS;
}
static int audio_open_output_stream(struct audio_hw_device *dev,
audio_io_handle_t handle,
audio_devices_t devices,
audio_output_flags_t flags,
struct audio_config *config,
struct audio_stream_out **stream_out)
{
struct a2dp_audio_dev *a2dp_dev = (struct a2dp_audio_dev *) dev;
struct a2dp_stream_out *out;
out = calloc(1, sizeof(struct a2dp_stream_out));
if (!out)
return -ENOMEM;
DBG("");
out->stream.common.get_sample_rate = out_get_sample_rate;
out->stream.common.set_sample_rate = out_set_sample_rate;
out->stream.common.get_buffer_size = out_get_buffer_size;
out->stream.common.get_channels = out_get_channels;
out->stream.common.get_format = out_get_format;
out->stream.common.set_format = out_set_format;
out->stream.common.standby = out_standby;
out->stream.common.dump = out_dump;
out->stream.common.set_parameters = out_set_parameters;
out->stream.common.get_parameters = out_get_parameters;
out->stream.common.add_audio_effect = out_add_audio_effect;
out->stream.common.remove_audio_effect = out_remove_audio_effect;
out->stream.get_latency = out_get_latency;
out->stream.set_volume = out_set_volume;
out->stream.write = out_write;
out->stream.get_render_position = out_get_render_position;
/* TODO: for now we always use endpoint 0 */
out->ep = &audio_endpoints[0];
if (!open_endpoint(out->ep, &out->cfg))
goto fail;
DBG("rate=%d channels=%d format=%d", out->cfg.rate,
out->cfg.channels, out->cfg.format);
if (out->cfg.channels == AUDIO_CHANNEL_OUT_MONO) {
out->downmix_buf = malloc(FIXED_BUFFER_SIZE / 2);
if (!out->downmix_buf)
goto fail;
}
*stream_out = &out->stream;
a2dp_dev->out = out;
out->audio_state = AUDIO_A2DP_STATE_STANDBY;
return 0;
fail:
error("audio: cannot open output stream");
free(out);
*stream_out = NULL;
return -EIO;
}
static void audio_close_output_stream(struct audio_hw_device *dev,
struct audio_stream_out *stream)
{
struct a2dp_audio_dev *a2dp_dev = (struct a2dp_audio_dev *) dev;
struct a2dp_stream_out *out = (struct a2dp_stream_out *) stream;
DBG("");
close_endpoint(a2dp_dev->out->ep);
free(out->downmix_buf);
free(stream);
a2dp_dev->out = NULL;
}
static int audio_set_parameters(struct audio_hw_device *dev,
const char *kvpairs)
{
struct a2dp_audio_dev *a2dp_dev = (struct a2dp_audio_dev *) dev;
struct a2dp_stream_out *out = a2dp_dev->out;
DBG("");
if (!out)
return 0;
return out->stream.common.set_parameters((struct audio_stream *) out,
kvpairs);
}
static char *audio_get_parameters(const struct audio_hw_device *dev,
const char *keys)
{
DBG("");
return strdup("");
}
static int audio_init_check(const struct audio_hw_device *dev)
{
DBG("");
return 0;
}
static int audio_set_voice_volume(struct audio_hw_device *dev, float volume)
{
DBG("");
return -ENOSYS;
}
static int audio_set_master_volume(struct audio_hw_device *dev, float volume)
{
DBG("");
return -ENOSYS;
}
static int audio_set_mode(struct audio_hw_device *dev, int mode)
{
DBG("");
return -ENOSYS;
}
static int audio_set_mic_mute(struct audio_hw_device *dev, bool state)
{
DBG("");
return -ENOSYS;
}
static int audio_get_mic_mute(const struct audio_hw_device *dev, bool *state)
{
DBG("");
return -ENOSYS;
}
static size_t audio_get_input_buffer_size(const struct audio_hw_device *dev,
const struct audio_config *config)
{
DBG("");
return -ENOSYS;
}
static int audio_open_input_stream(struct audio_hw_device *dev,
audio_io_handle_t handle,
audio_devices_t devices,
struct audio_config *config,
struct audio_stream_in **stream_in)
{
struct audio_stream_in *in;
DBG("");
in = calloc(1, sizeof(struct audio_stream_in));
if (!in)
return -ENOMEM;
in->common.get_sample_rate = in_get_sample_rate;
in->common.set_sample_rate = in_set_sample_rate;
in->common.get_buffer_size = in_get_buffer_size;
in->common.get_channels = in_get_channels;
in->common.get_format = in_get_format;
in->common.set_format = in_set_format;
in->common.standby = in_standby;
in->common.dump = in_dump;
in->common.set_parameters = in_set_parameters;
in->common.get_parameters = in_get_parameters;
in->common.add_audio_effect = in_add_audio_effect;
in->common.remove_audio_effect = in_remove_audio_effect;
in->set_gain = in_set_gain;
in->read = in_read;
in->get_input_frames_lost = in_get_input_frames_lost;
*stream_in = in;
return 0;
}
static void audio_close_input_stream(struct audio_hw_device *dev,
struct audio_stream_in *stream_in)
{
DBG("");
free(stream_in);
}
static int audio_dump(const audio_hw_device_t *device, int fd)
{
DBG("");
return -ENOSYS;
}
static int audio_close(hw_device_t *device)
{
struct a2dp_audio_dev *a2dp_dev = (struct a2dp_audio_dev *)device;
DBG("");
unregister_endpoints();
shutdown(listen_sk, SHUT_RDWR);
shutdown(audio_sk, SHUT_RDWR);
pthread_join(ipc_th, NULL);
close(listen_sk);
listen_sk = -1;
free(a2dp_dev);
return 0;
}
static void *ipc_handler(void *data)
{
bool done = false;
struct pollfd pfd;
int sk;
DBG("");
while (!done) {
DBG("Waiting for connection ...");
sk = accept(listen_sk, NULL, NULL);
if (sk < 0) {
int err = errno;
if (err == EINTR)
continue;
if (err != ECONNABORTED && err != EINVAL)
error("audio: Failed to accept socket: %d (%s)",
err, strerror(err));
break;
}
pthread_mutex_lock(&sk_mutex);
audio_sk = sk;
pthread_mutex_unlock(&sk_mutex);
DBG("Audio IPC: Connected");
if (register_endpoints() != AUDIO_STATUS_SUCCESS) {
error("audio: Failed to register endpoints");
unregister_endpoints();
pthread_mutex_lock(&sk_mutex);
shutdown(audio_sk, SHUT_RDWR);
close(audio_sk);
audio_sk = -1;
pthread_mutex_unlock(&sk_mutex);
continue;
}
memset(&pfd, 0, sizeof(pfd));
pfd.fd = audio_sk;
pfd.events = POLLHUP | POLLERR | POLLNVAL;
/* Check if socket is still alive. Empty while loop.*/
while (poll(&pfd, 1, -1) < 0 && errno == EINTR);
if (pfd.revents & (POLLHUP | POLLERR | POLLNVAL)) {
info("Audio HAL: Socket closed");
pthread_mutex_lock(&sk_mutex);
close(audio_sk);
audio_sk = -1;
pthread_mutex_unlock(&sk_mutex);
}
}
/* audio_sk is closed at this point, just cleanup endpoints states */
memset(audio_endpoints, 0, sizeof(audio_endpoints));
info("Closing Audio IPC thread");
return NULL;
}
static int audio_ipc_init(void)
{
struct sockaddr_un addr;
int err;
int sk;
DBG("");
sk = socket(PF_LOCAL, SOCK_SEQPACKET, 0);
if (sk < 0) {
err = -errno;
error("audio: Failed to create socket: %d (%s)", -err,
strerror(-err));
return err;
}
memset(&addr, 0, sizeof(addr));
addr.sun_family = AF_UNIX;
memcpy(addr.sun_path, BLUEZ_AUDIO_SK_PATH,
sizeof(BLUEZ_AUDIO_SK_PATH));
if (bind(sk, (struct sockaddr *) &addr, sizeof(addr)) < 0) {
err = -errno;
error("audio: Failed to bind socket: %d (%s)", -err,
strerror(-err));
goto failed;
}
if (listen(sk, 1) < 0) {
err = -errno;
error("audio: Failed to listen on the socket: %d (%s)", -err,
strerror(-err));
goto failed;
}
listen_sk = sk;
err = pthread_create(&ipc_th, NULL, ipc_handler, NULL);
if (err) {
err = -err;
ipc_th = 0;
error("audio: Failed to start Audio IPC thread: %d (%s)",
-err, strerror(-err));
goto failed;
}
return 0;
failed:
close(sk);
return err;
}
static int audio_open(const hw_module_t *module, const char *name,
hw_device_t **device)
{
struct a2dp_audio_dev *a2dp_dev;
int err;
DBG("");
if (strcmp(name, AUDIO_HARDWARE_INTERFACE)) {
error("audio: interface %s not matching [%s]", name,
AUDIO_HARDWARE_INTERFACE);
return -EINVAL;
}
err = audio_ipc_init();
if (err < 0)
return err;
a2dp_dev = calloc(1, sizeof(struct a2dp_audio_dev));
if (!a2dp_dev)
return -ENOMEM;
a2dp_dev->dev.common.tag = HARDWARE_DEVICE_TAG;
a2dp_dev->dev.common.version = AUDIO_DEVICE_API_VERSION_CURRENT;
a2dp_dev->dev.common.module = (struct hw_module_t *) module;
a2dp_dev->dev.common.close = audio_close;
a2dp_dev->dev.init_check = audio_init_check;
a2dp_dev->dev.set_voice_volume = audio_set_voice_volume;
a2dp_dev->dev.set_master_volume = audio_set_master_volume;
a2dp_dev->dev.set_mode = audio_set_mode;
a2dp_dev->dev.set_mic_mute = audio_set_mic_mute;
a2dp_dev->dev.get_mic_mute = audio_get_mic_mute;
a2dp_dev->dev.set_parameters = audio_set_parameters;
a2dp_dev->dev.get_parameters = audio_get_parameters;
a2dp_dev->dev.get_input_buffer_size = audio_get_input_buffer_size;
a2dp_dev->dev.open_output_stream = audio_open_output_stream;
a2dp_dev->dev.close_output_stream = audio_close_output_stream;
a2dp_dev->dev.open_input_stream = audio_open_input_stream;
a2dp_dev->dev.close_input_stream = audio_close_input_stream;
a2dp_dev->dev.dump = audio_dump;
/*
* Note that &a2dp_dev->dev.common is the same pointer as a2dp_dev.
* This results from the structure of following structs:a2dp_audio_dev,
* audio_hw_device. We will rely on this later in the code.
*/
*device = &a2dp_dev->dev.common;
return 0;
}
static struct hw_module_methods_t hal_module_methods = {
.open = audio_open,
};
struct audio_module HAL_MODULE_INFO_SYM = {
.common = {
.tag = HARDWARE_MODULE_TAG,
.version_major = 1,
.version_minor = 0,
.id = AUDIO_HARDWARE_MODULE_ID,
.name = "A2DP Bluez HW HAL",
.author = "Intel Corporation",
.methods = &hal_module_methods,
},
};