1
0
mirror of https://github.com/f4exb/sdrangel.git synced 2024-12-23 10:05:46 -05:00
sdrangel/plugins/samplemimo/xtrxmimo/xtrxmothread.cpp
2020-11-11 04:06:42 +01:00

240 lines
7.0 KiB
C++

///////////////////////////////////////////////////////////////////////////////////
// Copyright (C) 2020 Edouard Griffiths, F4EXB //
// //
// This program is free software; you can redistribute it and/or modify //
// it under the terms of the GNU General Public License as published by //
// the Free Software Foundation as version 3 of the License, or //
// (at your option) any later version. //
// //
// This program is distributed in the hope that it will be useful, //
// but WITHOUT ANY WARRANTY; without even the implied warranty of //
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the //
// GNU General Public License V3 for more details. //
// //
// You should have received a copy of the GNU General Public License //
// along with this program. If not, see <http://www.gnu.org/licenses/>. //
///////////////////////////////////////////////////////////////////////////////////
#include <algorithm>
#include <chrono>
#include <thread>
#include "dsp/samplemofifo.h"
#include "xtrxmothread.h"
XTRXMOThread::XTRXMOThread(struct xtrx_dev *dev, QObject* parent) :
QThread(parent),
m_running(false),
m_dev(dev),
m_sampleFifo(nullptr)
{
qDebug("XTRXMOThread::XTRXMOThread");
m_buf = new qint16[2*DeviceXTRX::blockSize*2];
std::fill(m_buf, m_buf + 2*DeviceXTRX::blockSize*2, 0);
}
XTRXMOThread::~XTRXMOThread()
{
qDebug("XTRXMOThread::~XTRXMOThread");
if (m_running) {
stopWork();
}
delete[] m_buf;
}
void XTRXMOThread::startWork()
{
m_startWaitMutex.lock();
start();
while(!m_running) {
m_startWaiter.wait(&m_startWaitMutex, 100);
}
m_startWaitMutex.unlock();
}
void XTRXMOThread::stopWork()
{
m_running = false;
wait();
}
void XTRXMOThread::setLog2Interpolation(unsigned int log2Interp)
{
qDebug("XTRXMOThread::setLog2Interpolation: %u", log2Interp);
m_log2Interp = log2Interp;
}
unsigned int XTRXMOThread::getLog2Interpolation() const
{
return m_log2Interp;
}
void XTRXMOThread::run()
{
int res;
m_running = true;
m_startWaiter.wakeAll();
xtrx_run_params params;
xtrx_run_params_init(&params);
params.dir = XTRX_TX;
params.tx_repeat_buf = 0;
params.tx.paketsize = 2*DeviceXTRX::blockSize;
params.tx.chs = XTRX_CH_AB;
params.tx.wfmt = XTRX_WF_16;
params.tx.hfmt = XTRX_IQ_INT16;
params.tx.flags |= XTRX_RSP_SWAP_IQ;
res = xtrx_run_ex(m_dev, &params);
if (res != 0)
{
qCritical("XTRXMOThread::run: could not start stream err:%d", res);
m_running = false;
}
else
{
std::this_thread::sleep_for(std::chrono::milliseconds(50));
qDebug("XTRXMOThread::run: stream started");
}
qint16 buf0[2*DeviceXTRX::blockSize]; // I+Q = 2x16 bit samples
qint16 buf1[2*DeviceXTRX::blockSize];
std::vector<void *> buffs(2);
master_ts ts = 4096*1024;
buffs[0] = &buf0;
buffs[1] = &buf1;
xtrx_send_ex_info_t nfo;
nfo.samples = DeviceXTRX::blockSize;
nfo.buffer_count = 2;
nfo.buffers = (void* const*) buffs.data();
nfo.flags = XTRX_TX_DONT_BUFFER; // | XTRX_TX_SEND_ZEROS;
nfo.timeout = 0;
nfo.out_txlatets = 0;
nfo.ts = ts;
while (m_running)
{
callback(buf0, buf1, nfo.samples);
res = xtrx_send_sync_ex(m_dev, &nfo);
if (res < 0)
{
qCritical("XTRXMOThread::run send error: %d", res);
qDebug("XTRXMOThread::run: out_samples: %u out_flags: %u", nfo.out_samples, nfo.out_flags);
break;
}
if (nfo.out_flags & XTRX_TX_DISCARDED_TO) {
qDebug("XTRXMOThread::run: underrun");
}
if (nfo.out_txlatets) {
qDebug("XTRXMOThread::run: out_txlatets: %lu", nfo.out_txlatets);
}
nfo.ts += DeviceXTRX::blockSize;
}
res = xtrx_stop(m_dev, XTRX_TX);
if (res != 0)
{
qCritical("XTRXMOThread::run: could not stop stream");
}
else
{
std::this_thread::sleep_for(std::chrono::milliseconds(50));
qDebug("XTRXMOThread::run: stream stopped");
}
m_running = false;
}
void XTRXMOThread::callback(qint16* buf0, qint16* buf1, qint32 samplesPerChannel)
{
unsigned int iPart1Begin, iPart1End, iPart2Begin, iPart2End;
m_sampleFifo->readSync(samplesPerChannel/(1<<m_log2Interp), iPart1Begin, iPart1End, iPart2Begin, iPart2End);
if (iPart1Begin != iPart1End)
{
callbackPart(buf0, buf1, (iPart1End - iPart1Begin)*(1<<m_log2Interp), iPart1Begin);
}
if (iPart2Begin != iPart2End)
{
unsigned int shift = (iPart1End - iPart1Begin)*(1<<m_log2Interp);
callbackPart(buf0 + 2*shift, buf1 + 2*shift, (iPart2End - iPart2Begin)*(1<<m_log2Interp), iPart2Begin);
}
}
// Interpolate according to specified log2 (ex: log2=4 => decim=16). len is a number of samples (not a number of I or Q)
void XTRXMOThread::callbackPart(qint16* buf0, qint16* buf1, qint32 nSamples, int iBegin)
{
for (unsigned int channel = 0; channel < 2; channel++)
{
SampleVector::iterator begin = m_sampleFifo->getData(channel).begin() + iBegin;
if (m_log2Interp == 0)
{
m_interpolators[channel].interpolate1(
&begin,
channel == 0 ? buf0 : buf1,
2*nSamples);
}
else
{
switch (m_log2Interp)
{
case 1:
m_interpolators[channel].interpolate2_cen(
&begin,
channel == 0 ? buf0 : buf1,
2*nSamples);
break;
case 2:
m_interpolators[channel].interpolate4_cen(
&begin,
channel == 0 ? buf0 : buf1,
2*nSamples);
break;
case 3:
m_interpolators[channel].interpolate8_cen(
&begin,
channel == 0 ? buf0 : buf1,
2*nSamples);
break;
case 4:
m_interpolators[channel].interpolate16_cen(
&begin,
channel == 0 ? buf0 : buf1,
2*nSamples);
break;
case 5:
m_interpolators[channel].interpolate32_cen(
&begin,
channel == 0 ? buf0 : buf1,
2*nSamples);
break;
case 6:
m_interpolators[channel].interpolate64_cen(
&begin,
channel == 0 ? buf0 : buf1,
2*nSamples);
break;
default:
break;
}
}
}
}