blob: 4a9e1446272c96d08841ecfd7d34507b4168438a [file] [log] [blame]
/*
* Copyright (C) 2018 The Android Open Source Project
*
* 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 <memory.h>
#define LOG_TAG "EffectHAL"
#define ATRACE_TAG ATRACE_TAG_AUDIO
#include "Effect.h"
#include "common/all-versions/default/EffectMap.h"
#define ATRACE_TAG ATRACE_TAG_AUDIO
#include <HidlUtils.h>
#include <android/log.h>
#include <cutils/properties.h>
#include <media/EffectsFactoryApi.h>
#include <mediautils/ScopedStatistics.h>
#include <sys/syscall.h>
#include <system/audio_effects/effect_spatializer.h>
#include <util/EffectUtils.h>
#include <utils/Trace.h>
#include "VersionUtils.h"
namespace android {
namespace hardware {
namespace audio {
namespace effect {
namespace CPP_VERSION {
namespace implementation {
#if MAJOR_VERSION <= 6
using ::android::hardware::audio::common::COMMON_TYPES_CPP_VERSION::implementation::
AudioChannelBitfield;
#endif
using ::android::hardware::audio::common::COMMON_TYPES_CPP_VERSION::implementation::HidlUtils;
namespace {
/**
* Some basic scheduling tools.
*/
namespace scheduler {
int getCpu() {
return sched_getcpu();
}
uint64_t getAffinity(pid_t tid) {
cpu_set_t set;
CPU_ZERO_S(sizeof(set), &set);
if (sched_getaffinity(tid, sizeof(set), &set)) {
ALOGW("%s: for tid:%d returning 0, failed %s", __func__, tid, strerror(errno));
return 0;
}
const int count = CPU_COUNT_S(sizeof(set), &set);
uint64_t mask = 0;
for (int i = 0; i < CPU_SETSIZE; ++i) {
if (CPU_ISSET_S(i, sizeof(set), &set)) {
mask |= 1 << i;
}
}
ALOGV("%s: for tid:%d returning cpu count %d mask %llu", __func__, tid, count,
(unsigned long long)mask);
return mask;
}
status_t setAffinity(pid_t tid, uint64_t mask) {
cpu_set_t set;
CPU_ZERO_S(sizeof(set), &set);
for (uint64_t m = mask; m != 0;) {
uint64_t tz = __builtin_ctz(m);
CPU_SET_S(tz, sizeof(set), &set);
m &= ~(1 << tz);
}
if (sched_setaffinity(tid, sizeof(set), &set)) {
ALOGW("%s: for tid:%d setting cpu mask %llu failed %s", __func__, tid,
(unsigned long long)mask, strerror(errno));
return -errno;
}
ALOGV("%s: for tid:%d setting cpu mask %llu", __func__, tid, (unsigned long long)mask);
return OK;
}
__unused status_t setPriority(pid_t tid, int policy, int priority) {
struct sched_param param {
.sched_priority = priority,
};
if (sched_setscheduler(tid, policy, &param) != 0) {
ALOGW("%s: Cannot set FIFO priority for tid %d to policy %d priority %d %s", __func__, tid,
policy, priority, strerror(errno));
return -errno;
}
ALOGV("%s: Successfully set priority for tid %d to policy %d priority %d", __func__, tid,
policy, priority);
return NO_ERROR;
}
status_t setUtilMin(pid_t tid, uint32_t utilMin) {
// Currently, there is no wrapper in bionic: b/183240349.
struct {
uint32_t size;
uint32_t sched_policy;
uint64_t sched_flags;
int32_t sched_nice;
uint32_t sched_priority;
uint64_t sched_runtime;
uint64_t sched_deadline;
uint64_t sched_period;
uint32_t sched_util_min;
uint32_t sched_util_max;
} attr{
.size = sizeof(attr),
.sched_flags = SCHED_FLAG_KEEP_ALL | SCHED_FLAG_UTIL_CLAMP_MIN,
.sched_util_min = utilMin,
};
if (syscall(__NR_sched_setattr, tid, &attr, 0 /* flags */)) {
ALOGW("%s: Cannot set sched_util_min for pid %d to %u %s", __func__, tid, utilMin,
strerror(errno));
return -errno;
}
ALOGV("%s: Successfully set sched_util_min for pid %d to %u", __func__, tid, utilMin);
return NO_ERROR;
}
/*
Attempts to raise the priority and usage of tid for spatialization.
Returns OK if everything works.
*/
status_t updateSpatializerPriority(pid_t tid) {
status_t status = OK;
const int cpu = getCpu();
ALOGV("%s: current CPU:%d", __func__, cpu);
const auto currentAffinity = getAffinity(tid);
ALOGV("%s: current Affinity:%llx", __func__, (unsigned long long)currentAffinity);
// Set the desired CPU core affinity.
// Typically this would be done to move the Spatializer effect off of the little cores.
// The mid cores and large cores typically have more FP/NEON units
// and will advantageously reduce power and prevent glitches due CPU limitations.
//
// Since this is SOC dependent, we do not set the core affinity here but
// prefer to set the util_clamp_min below.
//
constexpr uint64_t kDefaultAffinity = 0;
const int32_t desiredAffinity =
property_get_int32("audio.spatializer.effect.affinity", kDefaultAffinity);
if (desiredAffinity != 0 && (desiredAffinity & ~currentAffinity) == 0) {
const status_t localStatus = setAffinity(tid, desiredAffinity);
status = status ? status : localStatus;
}
// Set the util_clamp_min.
// This is beneficial to reduce glitches when starting up, or due to scheduler
// thread statistics reset (e.g. core migration), which cause the CPU frequency to drop
// to minimum.
//
// Experimentation has found that moving to a mid core over a little core reduces
// power if the mid core (e.g. A76/78) has more (e.g. 2x) FP/NEON units
// than the little core (e.g. A55).
// A possible value is 300.
//
constexpr uint32_t kUtilMin = 0;
const int32_t utilMin = property_get_int32("audio.spatializer.effect.util_clamp_min", kUtilMin);
if (utilMin > 0 && utilMin <= 1024) {
const status_t localStatus = setUtilMin(tid, utilMin);
status = status ? status : localStatus;
}
#if 0
// Provided for local vendor testing but not enabled as audioserver does this for us.
//
// Set priority if specified.
constexpr int32_t kRTPriorityMin = 1;
constexpr int32_t kRTPriorityMax = 3;
const int32_t priorityBoost =
property_get_int32("audio.spatializer.priority", kRTPriorityMin);
if (priorityBoost >= kRTPriorityMin && priorityBoost <= kRTPriorityMax) {
const status_t localStatus = scheduler::setPriority(threadId, SCHED_FIFO, priorityBoost);
status = status ? status : localStatus;
}
#endif
return status;
}
} // namespace scheduler
#define SCOPED_STATS() \
::android::mediautils::ScopedStatistics scopedStatistics { \
std::string("EffectHal::").append(__func__), mEffectHal->mStatistics \
}
class ProcessThread : public Thread {
public:
// ProcessThread's lifespan never exceeds Effect's lifespan.
ProcessThread(std::atomic<bool>* stop, effect_handle_t effect,
std::atomic<audio_buffer_t*>* inBuffer, std::atomic<audio_buffer_t*>* outBuffer,
Effect::StatusMQ* statusMQ, EventFlag* efGroup, Effect* effectHal)
: Thread(false /*canCallJava*/),
mStop(stop),
mEffect(effect),
mHasProcessReverse((*mEffect)->process_reverse != NULL),
mInBuffer(inBuffer),
mOutBuffer(outBuffer),
mStatusMQ(statusMQ),
mEfGroup(efGroup),
mEffectHal(effectHal) {}
virtual ~ProcessThread() {}
private:
std::atomic<bool>* mStop;
effect_handle_t mEffect;
bool mHasProcessReverse;
std::atomic<audio_buffer_t*>* mInBuffer;
std::atomic<audio_buffer_t*>* mOutBuffer;
Effect::StatusMQ* mStatusMQ;
EventFlag* mEfGroup;
Effect* const mEffectHal;
bool threadLoop() override;
};
bool ProcessThread::threadLoop() {
// This implementation doesn't return control back to the Thread until it decides to stop,
// as the Thread uses mutexes, and this can lead to priority inversion.
while (!std::atomic_load_explicit(mStop, std::memory_order_acquire)) {
uint32_t efState = 0;
mEfGroup->wait(static_cast<uint32_t>(MessageQueueFlagBits::REQUEST_PROCESS_ALL), &efState);
if (!(efState & static_cast<uint32_t>(MessageQueueFlagBits::REQUEST_PROCESS_ALL)) ||
(efState & static_cast<uint32_t>(MessageQueueFlagBits::REQUEST_QUIT))) {
continue; // Nothing to do or time to quit.
}
Result retval = Result::OK;
if (efState & static_cast<uint32_t>(MessageQueueFlagBits::REQUEST_PROCESS_REVERSE) &&
!mHasProcessReverse) {
retval = Result::NOT_SUPPORTED;
}
if (retval == Result::OK) {
// affects both buffer pointers and their contents.
std::atomic_thread_fence(std::memory_order_acquire);
int32_t processResult;
audio_buffer_t* inBuffer =
std::atomic_load_explicit(mInBuffer, std::memory_order_relaxed);
audio_buffer_t* outBuffer =
std::atomic_load_explicit(mOutBuffer, std::memory_order_relaxed);
if (inBuffer != nullptr && outBuffer != nullptr) {
// Time this effect process
SCOPED_STATS();
if (efState & static_cast<uint32_t>(MessageQueueFlagBits::REQUEST_PROCESS)) {
processResult = (*mEffect)->process(mEffect, inBuffer, outBuffer);
} else {
processResult = (*mEffect)->process_reverse(mEffect, inBuffer, outBuffer);
}
std::atomic_thread_fence(std::memory_order_release);
} else {
ALOGE("processing buffers were not set before calling 'process'");
processResult = -ENODEV;
}
switch (processResult) {
case 0:
retval = Result::OK;
break;
case -ENODATA:
retval = Result::INVALID_STATE;
break;
case -EINVAL:
retval = Result::INVALID_ARGUMENTS;
break;
default:
retval = Result::NOT_INITIALIZED;
}
}
if (!mStatusMQ->write(&retval)) {
ALOGW("status message queue write failed");
}
mEfGroup->wake(static_cast<uint32_t>(MessageQueueFlagBits::DONE_PROCESSING));
}
return false;
}
} // namespace
// static
const char* Effect::sContextResultOfCommand = "returned status";
const char* Effect::sContextCallToCommand = "error";
const char* Effect::sContextCallFunction = sContextCallToCommand;
const char* Effect::sContextConversion = "conversion";
Effect::Effect(bool isInput, effect_handle_t handle)
: mIsInput(isInput), mHandle(handle), mEfGroup(nullptr), mStopProcessThread(false) {
(void)mIsInput; // prevent 'unused field' warnings in pre-V7 versions.
}
Effect::~Effect() {
ATRACE_CALL();
auto [_, handle] = closeImpl();
if (mProcessThread.get()) {
ATRACE_NAME("mProcessThread->join");
status_t status = mProcessThread->join();
ALOGE_IF(status, "processing thread exit error: %s", strerror(-status));
}
if (mEfGroup) {
status_t status = EventFlag::deleteEventFlag(&mEfGroup);
ALOGE_IF(status, "processing MQ event flag deletion error: %s", strerror(-status));
}
mInBuffer.clear();
mOutBuffer.clear();
#if MAJOR_VERSION <= 5
int status = EffectRelease(handle);
ALOGW_IF(status, "Error releasing effect %p: %s", handle, strerror(-status));
#endif
EffectMap::getInstance().remove(handle);
}
// static
template <typename T>
size_t Effect::alignedSizeIn(size_t s) {
return (s + sizeof(T) - 1) / sizeof(T);
}
// static
template <typename T>
std::unique_ptr<uint8_t[]> Effect::hidlVecToHal(const hidl_vec<T>& vec, uint32_t* halDataSize) {
// Due to bugs in HAL, they may attempt to write into the provided
// input buffer. The original binder buffer is r/o, thus it is needed
// to create a r/w version.
*halDataSize = vec.size() * sizeof(T);
std::unique_ptr<uint8_t[]> halData(new uint8_t[*halDataSize]);
memcpy(&halData[0], &vec[0], *halDataSize);
return halData;
}
#if MAJOR_VERSION <= 6
void Effect::effectAuxChannelsConfigFromHal(const channel_config_t& halConfig,
EffectAuxChannelsConfig* config) {
config->mainChannels = AudioChannelBitfield(halConfig.main_channels);
config->auxChannels = AudioChannelBitfield(halConfig.aux_channels);
}
// static
void Effect::effectAuxChannelsConfigToHal(const EffectAuxChannelsConfig& config,
channel_config_t* halConfig) {
halConfig->main_channels = static_cast<audio_channel_mask_t>(config.mainChannels);
halConfig->aux_channels = static_cast<audio_channel_mask_t>(config.auxChannels);
}
#else // MAJOR_VERSION <= 6
void Effect::effectAuxChannelsConfigFromHal(const channel_config_t& halConfig,
EffectAuxChannelsConfig* config) {
(void)HidlUtils::audioChannelMaskFromHal(halConfig.main_channels, mIsInput,
&config->mainChannels);
(void)HidlUtils::audioChannelMaskFromHal(halConfig.aux_channels, mIsInput,
&config->auxChannels);
}
// static
void Effect::effectAuxChannelsConfigToHal(const EffectAuxChannelsConfig& config,
channel_config_t* halConfig) {
(void)HidlUtils::audioChannelMaskToHal(config.mainChannels, &halConfig->main_channels);
(void)HidlUtils::audioChannelMaskToHal(config.auxChannels, &halConfig->aux_channels);
}
#endif // MAJOR_VERSION <= 6
// static
void Effect::effectOffloadParamToHal(const EffectOffloadParameter& offload,
effect_offload_param_t* halOffload) {
halOffload->isOffload = offload.isOffload;
halOffload->ioHandle = offload.ioHandle;
}
// static
bool Effect::parameterToHal(uint32_t paramSize, const void* paramData, uint32_t valueSize,
const void** valueData, std::vector<uint8_t>* halParamBuffer) {
constexpr size_t kMaxSize = EFFECT_PARAM_SIZE_MAX - sizeof(effect_param_t);
if (paramSize > kMaxSize) {
ALOGE("%s: Parameter size is too big: %" PRIu32, __func__, paramSize);
return false;
}
size_t valueOffsetFromData = alignedSizeIn<uint32_t>(paramSize) * sizeof(uint32_t);
if (valueOffsetFromData > kMaxSize) {
ALOGE("%s: Aligned parameter size is too big: %zu", __func__, valueOffsetFromData);
return false;
}
if (valueSize > kMaxSize - valueOffsetFromData) {
ALOGE("%s: Value size is too big: %" PRIu32 ", max size is %zu", __func__, valueSize,
kMaxSize - valueOffsetFromData);
android_errorWriteLog(0x534e4554, "237291425");
return false;
}
size_t halParamBufferSize = sizeof(effect_param_t) + valueOffsetFromData + valueSize;
halParamBuffer->resize(halParamBufferSize, 0);
effect_param_t* halParam = reinterpret_cast<effect_param_t*>(halParamBuffer->data());
halParam->psize = paramSize;
halParam->vsize = valueSize;
memcpy(halParam->data, paramData, paramSize);
if (valueData) {
if (*valueData) {
// Value data is provided.
memcpy(halParam->data + valueOffsetFromData, *valueData, valueSize);
} else {
// The caller needs the pointer to the value data location.
*valueData = halParam->data + valueOffsetFromData;
}
}
return true;
}
Result Effect::analyzeCommandStatus(const char* commandName, const char* context, status_t status) {
return analyzeStatus("command", commandName, context, status);
}
Result Effect::analyzeStatus(const char* funcName, const char* subFuncName,
const char* contextDescription, status_t status) {
if (status != OK) {
ALOGW("Effect %p %s %s %s: %s", mHandle, funcName, subFuncName, contextDescription,
strerror(-status));
}
switch (status) {
case OK:
return Result::OK;
case -EINVAL:
return Result::INVALID_ARGUMENTS;
case -ENODATA:
return Result::INVALID_STATE;
case -ENODEV:
return Result::NOT_INITIALIZED;
case -ENOMEM:
return Result::RESULT_TOO_BIG;
case -ENOSYS:
return Result::NOT_SUPPORTED;
default:
return Result::INVALID_STATE;
}
}
#define RETURN_IF_EFFECT_CLOSED() \
if (mHandle == kInvalidEffectHandle) { \
return Result::INVALID_STATE; \
}
#define RETURN_RESULT_IF_EFFECT_CLOSED(result) \
if (mHandle == kInvalidEffectHandle) { \
_hidl_cb(Result::INVALID_STATE, result); \
return Void(); \
}
Return<void> Effect::getConfigImpl(int commandCode, const char* commandName,
GetConfigCallback _hidl_cb) {
RETURN_RESULT_IF_EFFECT_CLOSED(EffectConfig());
uint32_t halResultSize = sizeof(effect_config_t);
effect_config_t halConfig{};
status_t status =
(*mHandle)->command(mHandle, commandCode, 0, NULL, &halResultSize, &halConfig);
EffectConfig config;
if (status == OK) {
status = EffectUtils::effectConfigFromHal(halConfig, mIsInput, &config);
}
_hidl_cb(analyzeCommandStatus(commandName, sContextCallToCommand, status), config);
return Void();
}
Result Effect::getCurrentConfigImpl(uint32_t featureId, uint32_t configSize,
GetCurrentConfigSuccessCallback onSuccess) {
if (configSize > kMaxDataSize - sizeof(uint32_t)) {
ALOGE("%s: Config size is too big: %" PRIu32, __func__, configSize);
android_errorWriteLog(0x534e4554, "240266798");
return Result::INVALID_ARGUMENTS;
}
uint32_t halCmd = featureId;
std::vector<uint32_t> halResult(alignedSizeIn<uint32_t>(sizeof(uint32_t) + configSize), 0);
uint32_t halResultSize = 0;
return sendCommandReturningStatusAndData(
EFFECT_CMD_GET_FEATURE_CONFIG, "GET_FEATURE_CONFIG", sizeof(uint32_t), &halCmd,
&halResultSize, &halResult[0], sizeof(uint32_t), [&] { onSuccess(&halResult[1]); });
}
Result Effect::getParameterImpl(uint32_t paramSize, const void* paramData,
uint32_t requestValueSize, uint32_t replyValueSize,
GetParameterSuccessCallback onSuccess) {
// As it is unknown what method HAL uses for copying the provided parameter data,
// it is safer to make sure that input and output buffers do not overlap.
std::vector<uint8_t> halCmdBuffer;
if (!parameterToHal(paramSize, paramData, requestValueSize, nullptr, &halCmdBuffer)) {
return Result::INVALID_ARGUMENTS;
}
const void* valueData = nullptr;
std::vector<uint8_t> halParamBuffer;
if (!parameterToHal(paramSize, paramData, replyValueSize, &valueData, &halParamBuffer)) {
return Result::INVALID_ARGUMENTS;
}
uint32_t halParamBufferSize = halParamBuffer.size();
return sendCommandReturningStatusAndData(
EFFECT_CMD_GET_PARAM, "GET_PARAM", halCmdBuffer.size(), &halCmdBuffer[0],
&halParamBufferSize, &halParamBuffer[0], sizeof(effect_param_t), [&] {
effect_param_t* halParam = reinterpret_cast<effect_param_t*>(&halParamBuffer[0]);
onSuccess(halParam->vsize, valueData);
});
}
Result Effect::getSupportedConfigsImpl(uint32_t featureId, uint32_t maxConfigs, uint32_t configSize,
GetSupportedConfigsSuccessCallback onSuccess) {
if (maxConfigs != 0 && configSize > (kMaxDataSize - 2 * sizeof(uint32_t)) / maxConfigs) {
ALOGE("%s: Config size is too big: %" PRIu32, __func__, configSize);
return Result::INVALID_ARGUMENTS;
}
uint32_t halCmd[2] = {featureId, maxConfigs};
uint32_t halResultSize = 2 * sizeof(uint32_t) + maxConfigs * configSize;
std::vector<uint8_t> halResult(static_cast<size_t>(halResultSize), 0);
return sendCommandReturningStatusAndData(
EFFECT_CMD_GET_FEATURE_SUPPORTED_CONFIGS, "GET_FEATURE_SUPPORTED_CONFIGS", sizeof(halCmd),
halCmd, &halResultSize, &halResult[0], 2 * sizeof(uint32_t), [&] {
uint32_t* halResult32 = reinterpret_cast<uint32_t*>(&halResult[0]);
uint32_t supportedConfigs = *(++halResult32); // skip status field
if (supportedConfigs > maxConfigs) supportedConfigs = maxConfigs;
onSuccess(supportedConfigs, ++halResult32);
});
}
Return<void> Effect::prepareForProcessing(prepareForProcessing_cb _hidl_cb) {
RETURN_RESULT_IF_EFFECT_CLOSED(StatusMQ::Descriptor());
status_t status;
// Create message queue.
if (mStatusMQ) {
ALOGE("the client attempts to call prepareForProcessing_cb twice");
_hidl_cb(Result::INVALID_STATE, StatusMQ::Descriptor());
return Void();
}
std::unique_ptr<StatusMQ> tempStatusMQ(new StatusMQ(1, true /*EventFlag*/));
if (!tempStatusMQ->isValid()) {
ALOGE_IF(!tempStatusMQ->isValid(), "status MQ is invalid");
_hidl_cb(Result::INVALID_ARGUMENTS, StatusMQ::Descriptor());
return Void();
}
status = EventFlag::createEventFlag(tempStatusMQ->getEventFlagWord(), &mEfGroup);
if (status != OK || !mEfGroup) {
ALOGE("failed creating event flag for status MQ: %s", strerror(-status));
_hidl_cb(Result::INVALID_ARGUMENTS, StatusMQ::Descriptor());
return Void();
}
// Create and launch the thread.
mProcessThread = new ProcessThread(&mStopProcessThread, mHandle, &mHalInBufferPtr,
&mHalOutBufferPtr, tempStatusMQ.get(), mEfGroup, this);
status = mProcessThread->run("effect", PRIORITY_URGENT_AUDIO);
if (status != OK) {
ALOGW("failed to start effect processing thread: %s", strerror(-status));
_hidl_cb(Result::INVALID_ARGUMENTS, MQDescriptorSync<Result>());
return Void();
}
// For a spatializer effect, we perform scheduler adjustments to reduce glitches and power.
// We do it here instead of the ProcessThread::threadLoop to ensure that mHandle is valid.
if (effect_descriptor_t halDescriptor{};
(*mHandle)->get_descriptor(mHandle, &halDescriptor) == NO_ERROR &&
memcmp(&halDescriptor.type, FX_IID_SPATIALIZER, sizeof(effect_uuid_t)) == 0) {
const status_t status = scheduler::updateSpatializerPriority(mProcessThread->getTid());
ALOGW_IF(status != OK, "Failed to update Spatializer priority");
}
mStatusMQ = std::move(tempStatusMQ);
_hidl_cb(Result::OK, *mStatusMQ->getDesc());
return Void();
}
Return<Result> Effect::setProcessBuffers(const AudioBuffer& inBuffer,
const AudioBuffer& outBuffer) {
RETURN_IF_EFFECT_CLOSED();
AudioBufferManager& manager = AudioBufferManager::getInstance();
sp<AudioBufferWrapper> tempInBuffer, tempOutBuffer;
if (!manager.wrap(inBuffer, &tempInBuffer)) {
ALOGE("Could not map memory of the input buffer");
return Result::INVALID_ARGUMENTS;
}
if (!manager.wrap(outBuffer, &tempOutBuffer)) {
ALOGE("Could not map memory of the output buffer");
return Result::INVALID_ARGUMENTS;
}
mInBuffer = tempInBuffer;
mOutBuffer = tempOutBuffer;
// The processing thread only reads these pointers after waking up by an event flag,
// so it's OK to update the pair non-atomically.
mHalInBufferPtr.store(mInBuffer->getHalBuffer(), std::memory_order_release);
mHalOutBufferPtr.store(mOutBuffer->getHalBuffer(), std::memory_order_release);
return Result::OK;
}
Result Effect::sendCommand(int commandCode, const char* commandName) {
return sendCommand(commandCode, commandName, 0, NULL);
}
Result Effect::sendCommand(int commandCode, const char* commandName, uint32_t size, void* data) {
RETURN_IF_EFFECT_CLOSED();
status_t status = (*mHandle)->command(mHandle, commandCode, size, data, 0, NULL);
return analyzeCommandStatus(commandName, sContextCallToCommand, status);
}
Result Effect::sendCommandReturningData(int commandCode, const char* commandName,
uint32_t* replySize, void* replyData) {
return sendCommandReturningData(commandCode, commandName, 0, NULL, replySize, replyData);
}
Result Effect::sendCommandReturningData(int commandCode, const char* commandName, uint32_t size,
void* data, uint32_t* replySize, void* replyData) {
RETURN_IF_EFFECT_CLOSED();
uint32_t expectedReplySize = *replySize;
status_t status = (*mHandle)->command(mHandle, commandCode, size, data, replySize, replyData);
if (status == OK && *replySize != expectedReplySize) {
status = -ENODATA;
}
return analyzeCommandStatus(commandName, sContextCallToCommand, status);
}
Result Effect::sendCommandReturningStatus(int commandCode, const char* commandName) {
return sendCommandReturningStatus(commandCode, commandName, 0, NULL);
}
Result Effect::sendCommandReturningStatus(int commandCode, const char* commandName, uint32_t size,
void* data) {
uint32_t replyCmdStatus;
uint32_t replySize = sizeof(uint32_t);
return sendCommandReturningStatusAndData(commandCode, commandName, size, data, &replySize,
&replyCmdStatus, replySize, [] {});
}
Result Effect::sendCommandReturningStatusAndData(int commandCode, const char* commandName,
uint32_t size, void* data, uint32_t* replySize,
void* replyData, uint32_t minReplySize,
CommandSuccessCallback onSuccess) {
RETURN_IF_EFFECT_CLOSED();
status_t status = (*mHandle)->command(mHandle, commandCode, size, data, replySize, replyData);
Result retval;
if (status == OK && minReplySize >= sizeof(uint32_t) && *replySize >= minReplySize) {
uint32_t commandStatus = *reinterpret_cast<uint32_t*>(replyData);
retval = analyzeCommandStatus(commandName, sContextResultOfCommand, commandStatus);
if (commandStatus == OK) {
onSuccess();
}
} else {
retval = analyzeCommandStatus(commandName, sContextCallToCommand, status);
}
return retval;
}
Result Effect::setConfigImpl(int commandCode, const char* commandName, const EffectConfig& config,
const sp<IEffectBufferProviderCallback>& inputBufferProvider,
const sp<IEffectBufferProviderCallback>& outputBufferProvider) {
effect_config_t halConfig;
EffectUtils::effectConfigToHal(config, &halConfig);
if (inputBufferProvider != 0) {
LOG_FATAL("Using input buffer provider is not supported");
}
if (outputBufferProvider != 0) {
LOG_FATAL("Using output buffer provider is not supported");
}
return sendCommandReturningStatus(commandCode, commandName, sizeof(effect_config_t),
&halConfig);
}
Result Effect::setParameterImpl(uint32_t paramSize, const void* paramData, uint32_t valueSize,
const void* valueData) {
std::vector<uint8_t> halParamBuffer;
if (!parameterToHal(paramSize, paramData, valueSize, &valueData, &halParamBuffer)) {
return Result::INVALID_ARGUMENTS;
}
return sendCommandReturningStatus(EFFECT_CMD_SET_PARAM, "SET_PARAM", halParamBuffer.size(),
&halParamBuffer[0]);
}
// Methods from ::android::hardware::audio::effect::CPP_VERSION::IEffect follow.
Return<Result> Effect::init() {
return sendCommandReturningStatus(EFFECT_CMD_INIT, "INIT");
}
Return<Result> Effect::setConfig(const EffectConfig& config,
const sp<IEffectBufferProviderCallback>& inputBufferProvider,
const sp<IEffectBufferProviderCallback>& outputBufferProvider) {
return setConfigImpl(EFFECT_CMD_SET_CONFIG, "SET_CONFIG", config, inputBufferProvider,
outputBufferProvider);
}
Return<Result> Effect::reset() {
return sendCommand(EFFECT_CMD_RESET, "RESET");
}
Return<Result> Effect::enable() {
return sendCommandReturningStatus(EFFECT_CMD_ENABLE, "ENABLE");
}
Return<Result> Effect::disable() {
return sendCommandReturningStatus(EFFECT_CMD_DISABLE, "DISABLE");
}
Return<Result> Effect::setAudioSource(
#if MAJOR_VERSION <= 6
AudioSource source
#else
const AudioSource& source
#endif
) {
audio_source_t halSource;
if (status_t status = HidlUtils::audioSourceToHal(source, &halSource); status == NO_ERROR) {
uint32_t halSourceParam = static_cast<uint32_t>(halSource);
return sendCommand(EFFECT_CMD_SET_AUDIO_SOURCE, "SET_AUDIO_SOURCE", sizeof(uint32_t),
&halSourceParam);
} else {
return analyzeStatus(__func__, "audioSourceToHal", sContextConversion, status);
}
}
#if MAJOR_VERSION <= 6
Return<Result> Effect::setDevice(AudioDeviceBitfield device) {
uint32_t halDevice = static_cast<uint32_t>(device);
return sendCommand(EFFECT_CMD_SET_DEVICE, "SET_DEVICE", sizeof(uint32_t), &halDevice);
}
Return<Result> Effect::setInputDevice(AudioDeviceBitfield device) {
uint32_t halDevice = static_cast<uint32_t>(device);
return sendCommand(EFFECT_CMD_SET_INPUT_DEVICE, "SET_INPUT_DEVICE", sizeof(uint32_t),
&halDevice);
}
#else // MAJOR_VERSION <= 6
Return<Result> Effect::setDevice(const DeviceAddress& device) {
audio_devices_t halDevice;
char halDeviceAddress[AUDIO_DEVICE_MAX_ADDRESS_LEN];
if (status_t status = HidlUtils::deviceAddressToHal(device, &halDevice, halDeviceAddress);
status == NO_ERROR) {
uint32_t halDeviceParam = static_cast<uint32_t>(halDevice);
return sendCommand(EFFECT_CMD_SET_DEVICE, "SET_DEVICE", sizeof(uint32_t), &halDeviceParam);
} else {
return analyzeStatus(__func__, "deviceAddressToHal", sContextConversion, status);
}
}
Return<Result> Effect::setInputDevice(const DeviceAddress& device) {
audio_devices_t halDevice;
char halDeviceAddress[AUDIO_DEVICE_MAX_ADDRESS_LEN];
if (status_t status = HidlUtils::deviceAddressToHal(device, &halDevice, halDeviceAddress);
status == NO_ERROR) {
uint32_t halDeviceParam = static_cast<uint32_t>(halDevice);
return sendCommand(EFFECT_CMD_SET_INPUT_DEVICE, "SET_INPUT_DEVICE", sizeof(uint32_t),
&halDeviceParam);
} else {
return analyzeStatus(__func__, "deviceAddressToHal", sContextConversion, status);
}
}
#endif // MAJOR_VERSION <= 6
Return<void> Effect::setAndGetVolume(const hidl_vec<uint32_t>& volumes,
setAndGetVolume_cb _hidl_cb) {
uint32_t halDataSize;
std::unique_ptr<uint8_t[]> halData = hidlVecToHal(volumes, &halDataSize);
uint32_t halResultSize = halDataSize;
std::vector<uint32_t> halResult(volumes.size(), 0);
Result retval = sendCommandReturningData(EFFECT_CMD_SET_VOLUME, "SET_VOLUME", halDataSize,
&halData[0], &halResultSize, &halResult[0]);
hidl_vec<uint32_t> result;
if (retval == Result::OK) {
result.setToExternal(&halResult[0], halResultSize);
}
_hidl_cb(retval, result);
return Void();
}
Return<Result> Effect::volumeChangeNotification(const hidl_vec<uint32_t>& volumes) {
uint32_t halDataSize;
std::unique_ptr<uint8_t[]> halData = hidlVecToHal(volumes, &halDataSize);
return sendCommand(EFFECT_CMD_SET_VOLUME, "SET_VOLUME", halDataSize, &halData[0]);
}
Return<Result> Effect::setAudioMode(AudioMode mode) {
uint32_t halMode = static_cast<uint32_t>(mode);
return sendCommand(EFFECT_CMD_SET_AUDIO_MODE, "SET_AUDIO_MODE", sizeof(uint32_t), &halMode);
}
Return<Result> Effect::setConfigReverse(
const EffectConfig& config, const sp<IEffectBufferProviderCallback>& inputBufferProvider,
const sp<IEffectBufferProviderCallback>& outputBufferProvider) {
return setConfigImpl(EFFECT_CMD_SET_CONFIG_REVERSE, "SET_CONFIG_REVERSE", config,
inputBufferProvider, outputBufferProvider);
}
Return<void> Effect::getConfig(getConfig_cb _hidl_cb) {
return getConfigImpl(EFFECT_CMD_GET_CONFIG, "GET_CONFIG", _hidl_cb);
}
Return<void> Effect::getConfigReverse(getConfigReverse_cb _hidl_cb) {
return getConfigImpl(EFFECT_CMD_GET_CONFIG_REVERSE, "GET_CONFIG_REVERSE", _hidl_cb);
}
Return<void> Effect::getSupportedAuxChannelsConfigs(uint32_t maxConfigs,
getSupportedAuxChannelsConfigs_cb _hidl_cb) {
hidl_vec<EffectAuxChannelsConfig> result;
Result retval = getSupportedConfigsImpl(
EFFECT_FEATURE_AUX_CHANNELS, maxConfigs, sizeof(channel_config_t),
[&](uint32_t supportedConfigs, void* configsData) {
result.resize(supportedConfigs);
channel_config_t* config = reinterpret_cast<channel_config_t*>(configsData);
for (size_t i = 0; i < result.size(); ++i) {
effectAuxChannelsConfigFromHal(*config++, &result[i]);
}
});
_hidl_cb(retval, result);
return Void();
}
Return<void> Effect::getAuxChannelsConfig(getAuxChannelsConfig_cb _hidl_cb) {
EffectAuxChannelsConfig result;
Result retval = getCurrentConfigImpl(
EFFECT_FEATURE_AUX_CHANNELS, sizeof(channel_config_t), [&](void* configData) {
effectAuxChannelsConfigFromHal(*reinterpret_cast<channel_config_t*>(configData),
&result);
});
_hidl_cb(retval, result);
return Void();
}
Return<Result> Effect::setAuxChannelsConfig(const EffectAuxChannelsConfig& config) {
std::vector<uint32_t> halCmd(
alignedSizeIn<uint32_t>(sizeof(uint32_t) + sizeof(channel_config_t)), 0);
halCmd[0] = EFFECT_FEATURE_AUX_CHANNELS;
effectAuxChannelsConfigToHal(config, reinterpret_cast<channel_config_t*>(&halCmd[1]));
return sendCommandReturningStatus(EFFECT_CMD_SET_FEATURE_CONFIG,
"SET_FEATURE_CONFIG AUX_CHANNELS", halCmd.size(), &halCmd[0]);
}
Return<Result> Effect::offload(const EffectOffloadParameter& param) {
effect_offload_param_t halParam;
effectOffloadParamToHal(param, &halParam);
return sendCommandReturningStatus(EFFECT_CMD_OFFLOAD, "OFFLOAD", sizeof(effect_offload_param_t),
&halParam);
}
Return<void> Effect::getDescriptor(getDescriptor_cb _hidl_cb) {
RETURN_RESULT_IF_EFFECT_CLOSED(EffectDescriptor());
effect_descriptor_t halDescriptor;
memset(&halDescriptor, 0, sizeof(effect_descriptor_t));
status_t status = (*mHandle)->get_descriptor(mHandle, &halDescriptor);
EffectDescriptor descriptor;
if (status == OK) {
status = EffectUtils::effectDescriptorFromHal(halDescriptor, &descriptor);
}
_hidl_cb(analyzeStatus("get_descriptor", "", sContextCallFunction, status), descriptor);
return Void();
}
Return<void> Effect::command(uint32_t commandId, const hidl_vec<uint8_t>& data,
uint32_t resultMaxSize, command_cb _hidl_cb) {
if (mHandle == kInvalidEffectHandle) {
_hidl_cb(-ENODATA, hidl_vec<uint8_t>());
return Void();
}
uint32_t halDataSize;
std::unique_ptr<uint8_t[]> halData = hidlVecToHal(data, &halDataSize);
uint32_t halResultSize = resultMaxSize;
std::unique_ptr<uint8_t[]> halResult(new uint8_t[halResultSize]);
memset(&halResult[0], 0, halResultSize);
void* dataPtr = halDataSize > 0 ? &halData[0] : NULL;
void* resultPtr = halResultSize > 0 ? &halResult[0] : NULL;
status_t status = BAD_VALUE;
switch (commandId) {
case 'gtid': // retrieve the tid, used for spatializer priority boost
if (halDataSize == 0 && resultMaxSize == sizeof(int32_t)) {
auto ptid = (int32_t*)resultPtr;
ptid[0] = mProcessThread ? mProcessThread->getTid() : -1;
status = OK;
break; // we have handled 'gtid' here.
}
[[fallthrough]]; // allow 'gtid' overload (checked halDataSize and resultMaxSize).
default:
status = (*mHandle)->command(mHandle, commandId, halDataSize, dataPtr, &halResultSize,
resultPtr);
break;
}
hidl_vec<uint8_t> result;
if (status == OK && resultPtr != NULL) {
result.setToExternal(&halResult[0], halResultSize);
}
_hidl_cb(status, result);
return Void();
}
Return<Result> Effect::setParameter(const hidl_vec<uint8_t>& parameter,
const hidl_vec<uint8_t>& value) {
return setParameterImpl(parameter.size(), &parameter[0], value.size(), &value[0]);
}
Return<void> Effect::getParameter(const hidl_vec<uint8_t>& parameter, uint32_t valueMaxSize,
getParameter_cb _hidl_cb) {
hidl_vec<uint8_t> value;
Result retval = getParameterImpl(
parameter.size(), &parameter[0], valueMaxSize,
[&](uint32_t valueSize, const void* valueData) {
value.setToExternal(reinterpret_cast<uint8_t*>(const_cast<void*>(valueData)),
valueSize);
});
_hidl_cb(retval, value);
return Void();
}
Return<void> Effect::getSupportedConfigsForFeature(uint32_t featureId, uint32_t maxConfigs,
uint32_t configSize,
getSupportedConfigsForFeature_cb _hidl_cb) {
uint32_t configCount = 0;
hidl_vec<uint8_t> result;
Result retval = getSupportedConfigsImpl(featureId, maxConfigs, configSize,
[&](uint32_t supportedConfigs, void* configsData) {
configCount = supportedConfigs;
result.resize(configCount * configSize);
memcpy(&result[0], configsData, result.size());
});
_hidl_cb(retval, configCount, result);
return Void();
}
Return<void> Effect::getCurrentConfigForFeature(uint32_t featureId, uint32_t configSize,
getCurrentConfigForFeature_cb _hidl_cb) {
hidl_vec<uint8_t> result;
Result retval = getCurrentConfigImpl(featureId, configSize, [&](void* configData) {
result.resize(configSize);
memcpy(&result[0], configData, result.size());
});
_hidl_cb(retval, result);
return Void();
}
Return<Result> Effect::setCurrentConfigForFeature(uint32_t featureId,
const hidl_vec<uint8_t>& configData) {
std::vector<uint32_t> halCmd(alignedSizeIn<uint32_t>(sizeof(uint32_t) + configData.size()), 0);
halCmd[0] = featureId;
memcpy(&halCmd[1], &configData[0], configData.size());
return sendCommandReturningStatus(EFFECT_CMD_SET_FEATURE_CONFIG, "SET_FEATURE_CONFIG",
halCmd.size(), &halCmd[0]);
}
std::tuple<Result, effect_handle_t> Effect::closeImpl() {
if (mStopProcessThread.load(std::memory_order_relaxed)) { // only this thread modifies
return {Result::INVALID_STATE, kInvalidEffectHandle};
}
mStopProcessThread.store(true, std::memory_order_release);
if (mEfGroup) {
mEfGroup->wake(static_cast<uint32_t>(MessageQueueFlagBits::REQUEST_QUIT));
}
effect_handle_t handle = mHandle;
mHandle = kInvalidEffectHandle;
#if MAJOR_VERSION <= 5
return {Result::OK, handle};
#elif MAJOR_VERSION >= 6
// No need to join the processing thread, it is part of the API contract that the client
// must finish processing before closing the effect.
Result retval = analyzeStatus("EffectRelease", "", sContextCallFunction, EffectRelease(handle));
EffectMap::getInstance().remove(handle);
return {retval, handle};
#endif
}
Return<Result> Effect::close() {
RETURN_IF_EFFECT_CLOSED();
auto [result, _] = closeImpl();
return result;
}
Return<void> Effect::debug(const hidl_handle& fd, const hidl_vec<hidl_string>& /* options */) {
if (fd.getNativeHandle() != nullptr && fd->numFds == 1) {
uint32_t cmdData = fd->data[0];
(void)sendCommand(EFFECT_CMD_DUMP, "DUMP", sizeof(cmdData), &cmdData);
const std::string s = mStatistics->dump();
if (s.size() != 0) write(cmdData, s.c_str(), s.size());
}
return Void();
}
} // namespace implementation
} // namespace CPP_VERSION
} // namespace effect
} // namespace audio
} // namespace hardware
} // namespace android