mirror of
https://github.com/cjcliffe/CubicSDR.git
synced 2024-11-26 21:58:37 -05:00
SoapySDRThread tweak: directly resize to the final size when building SDRThreadIQData
This commit is contained in:
parent
af20f680a4
commit
5797548739
@ -209,11 +209,14 @@ void SDRThread::readStream(SDRThreadIQDataQueue* iqDataOutQueue) {
|
||||
//0. Retreive a new batch
|
||||
SDRThreadIQData *dataOut = buffers.getBuffer();
|
||||
|
||||
//resize to the target size immedialetly, to minimize later reallocs:
|
||||
assureBufferMinSize(dataOut, nElems);
|
||||
|
||||
//1.If overflow occured on the previous readStream(), transfer it in dataOut directly.
|
||||
//Take care of the iq_swap option.
|
||||
if (numOverflow > 0) {
|
||||
int n_overflow = std::min(numOverflow, nElems);
|
||||
|
||||
//safety
|
||||
assureBufferMinSize(dataOut, n_overflow);
|
||||
|
||||
::memcpy(&dataOut->data[0], &overflowBuffer.data[0], n_overflow * sizeof(liquid_float_complex));
|
||||
@ -261,6 +264,7 @@ void SDRThread::readStream(SDRThreadIQDataQueue* iqDataOutQueue) {
|
||||
//nor that the Re/Im layout of fields matches the float array order, assign liquid_float_complex field by field.
|
||||
float *pp = (float *)buffs[0];
|
||||
|
||||
//safety
|
||||
assureBufferMinSize(dataOut, n_read + n_requested);
|
||||
|
||||
if (iq_swap.load()) {
|
||||
@ -286,6 +290,7 @@ void SDRThread::readStream(SDRThreadIQDataQueue* iqDataOutQueue) {
|
||||
// std::cout << "SDRThread::readStream(): 2. SoapySDR read make nElems overflow by " << numNewOverflow << " samples..." << std::endl;
|
||||
}
|
||||
|
||||
//safety
|
||||
assureBufferMinSize(&overflowBuffer, numOverflow + numNewOverflow);
|
||||
|
||||
if (iq_swap.load()) {
|
||||
@ -308,6 +313,7 @@ void SDRThread::readStream(SDRThreadIQDataQueue* iqDataOutQueue) {
|
||||
|
||||
float *pp = (float *)buffs[0];
|
||||
|
||||
//safety
|
||||
assureBufferMinSize(dataOut, n_read + n_stream_read);
|
||||
|
||||
if (iq_swap.load()) {
|
||||
@ -332,7 +338,7 @@ void SDRThread::readStream(SDRThreadIQDataQueue* iqDataOutQueue) {
|
||||
//3. At that point, dataOut contains nElems (or less if a read has return an error), try to post in queue, else discard.
|
||||
if (n_read > 0 && !stopping && !iqDataOutQueue->full()) {
|
||||
|
||||
//clamp result:
|
||||
//clamp result to the actual read size:
|
||||
dataOut->data.resize(n_read);
|
||||
|
||||
dataOut->frequency = frequency.load();
|
||||
@ -354,6 +360,8 @@ void SDRThread::readStream(SDRThreadIQDataQueue* iqDataOutQueue) {
|
||||
else {
|
||||
dataOut->setRefCount(0);
|
||||
std::cout << "SDRThread::readStream(): 3.1 iqDataOutQueue output queue is full, discard processing of the batch..." << std::endl;
|
||||
//saturation, let a chance to the other threads to consume the existing samples
|
||||
std::this_thread::yield();
|
||||
}
|
||||
}
|
||||
|
||||
|
Loading…
Reference in New Issue
Block a user