Tag/Release 0.2.0

This is the first real tagged Osmocom release of OsmoTRX.

  [ Alexander Chemeris ]
  * EDGE: Add support for UmTRX.
  * Common: Get rid of a compilation warning.
  * Common: Make sure gLogEarly() log to the same facilities as the normal log.
  * transceiver: Properly handle MAXDLY.
  * transceiver: Add an option to generate random Access Bursts.
  * osmo-trx: Output Rx SPS as a part of configuration output.
  * transceiver: Do not pass transceiver state struct to function where it's not used.
  * makefile: Fix build from an external path.
  * radioDevice: GSMRATE macro must have parentheses around its definition.
  * uhd: Fix comment.
  * radioInterface: Initialize power scale with a meaningful default.
  * transceiver: Log channel number in DEBUG output of demoded bursts.
  * transceiver: Add an option to emulate a RACH delay in random filler mode.
  * UHD: Initial LimeSDR support.
  * CommonLibs: Remove unused files.
  * sigProcLib: Typo sybols -> symbols
  * radioBuffer: Remove extra ; at the end of inline function definitions.
  * sigProcLib: Fix documentation, sync argument names in .cpp and .h files.
  * sigProcLib: make energyDetect() simpler by returning actual energy.
  * sigProcLib: Rename demodulateBurst() to demodGmskBurst() for clarity.
  * sigProcLib: Slice SoftVector instead of signalVector for GMSK demod.
  * Call vectorSlicer() right before packing bits for transmission to osmo-bts.
  * CommonLibs: Print soft bits with less confidence to console when printing a soft vector.
  * BitVector: Remove convolutional codec - we don't use it in osmo-trx.
  * BitVector: Convert SoftVector from 0..1 to -1..+1 soft bits.
  * signalVector: Implement segment().
  * vector: Introduce segmentMove() method to move data inside of a vector.
  * vector: Introduce shrink() function to shrink vector size without loosing data.
  * Move CorrType type from Transceiver to sigProcLib.
  * sigProcLib: rename signalError type to SignalError.
  * Move Transceiver::detectBurst() to sigProcLib to make it reusable.
  * Move BURST_THRESH from Transceiver.cpp to sigProcLib.h to make it reusable.
  * sigProcLib: Add operator<< to print CorrType to a string.
  * sigProcLib.h: Fix whitespaces. No non-whitespace changes.
  * Move Transceiver::demodulate() to sigProcLib to make it reusable.
  * sigProcLib: constify signalVector arguments for detectBurst() functions.
  * sigProcLib: Constify demodulation functions burst argument.
  * sigProcLib: Fix number of tail bits in random Normal Bursts and zero Stealing Bits.
  * Configuration: Variables allocated with 'new' must be freed with 'delete'.
  * BitVector: Remove Generator class.
  * PRBS: a Pseudo-random binary sequence (PRBS) generator class.

  [ Tom Tsou ]
  * EDGE: Fix USRP B210 device support
  * uhd: Correct timing alignment in 8-PSK and GMSK downlink bursts
  * EDGE: Fix demodulation slicer input
  * common: Restrict UDP binding to localhost only
  * common: Add mandatory length field to UDP receive calls
  * uhd: Update default E3XX settings
  * uhd: Set default Tx sampling to 4 sps
  * uhd: Make device offset check a private method
  * uhd: Set minimum UHD version requirement for E3XX
  * sigproc: Expand RACH, TSC, and EDGE correlation windows
  * transceiver: Do not report error on SETTSC when radio is on
  * transceiver: Add Rx samples-per-symbol option
  * radioInterface: Convert diversity argument to general type
  * iface: Add inner ring-buffer implementation
  * mcbts: Add multi-ARFCN channelizing filters
  * mcbts: Add multi-ARFCN radio support
  * sigproc: Adjust burst detection threshold criteria
  * egprs: Enable 8-PSK length vectors on the Tx interface
  * egprs: Enable 8-PSK burst detection when EDGE is enabled
  * transceiver: Remove HANDOVER warnings
  * mcbts: Allow out of order channel setup
  * radioInterface: Fix multi-channel buffer index bug
  * uhd: Add command line option for GPS reference
  * transceiver: Fix mixed GSMK / 8-PSK transmission
  * transceiver: Fix 4 SPS receive TOA value
  * sigproc: Fix missing 8-PSK tail symbols
  * uhd: Update USRP2/N200/N210 for 4 SPS Rx
  * sigproc: Match differential GMSK start/end bits to tail bits
  * uhd: Add missing B200 sample timing for 4 SPS receive
  * transceiver: Fix command build warning
  * uhd: Set minimum supported version to 3.9.0
  * uhd: Add X300 sample timing for 4 SPS
  * Revert "uhd: Set minimum supported version to 3.9.0"
  * uhd: Add support for UHD-3.11 logging control
  * uhd: Increase MC-BTS FPGA clock rate to 51.2 MHz
  * Resampler: Fix initialization return checking
  * sigProcLib: Remove unreachable code and no-effect checks
  * sigProcLib: Check return status on downsampling
  * sigProcLib: Fix negative value check on unsigned value
  * Resampler: Fix non-array delete for filter taps
  * Transceiver: Remove unsigned negative compares
  * Configuration: Fix const and signedness compile warnings
  * config: Remove OpenBTS style sqlite configuration
  * radioInterface: Remove UmTRX 'diversity' option
  * build: Require and check for gcc C++11 support
  * uhd: Use map container for for device parameter access
  * sigProcLib: Remove unused functions from public interface
  * uhd: Add non-UmTRX channel swap support
  * uhd: Fix Tx-RX timing offset setting
  * uhd: Fix USRP2/N200/N210 device detection
  * transceiver: Fix POWEROFF crash on USRP2/N200/X300 devices
  * sigProcLib: Fix complex/real vector flag in Laurent modulator
  * sigProcLib: Remove heap based signal vector allocations
  * common: Declare explicit Vector move constructor
  * sigProcLib: Remove trigonometric tables
  * sigProcLib: Use explicit NaN check in sinc table generation
  * sigProcLib: Replace dynamically allocated resampling buffers
  * sigProcLib: Specify standard namespace for isnan()
  * uhd: Always specify samples-per-symbol for device lookup
  * LimeSDR: set approximate tx offset value to make GSM work

  [ Neels Hofmeyr ]
  * add basic .gitignore
  * configure.ac: check for boost/config.hpp header
  * The INSTALL file is being overwritten by autoreconf, but it is committed as empty file. As a result, the INSTALL file always shows as modified. Instead, remove INSTALL from git and ignore it.
  * add contrib/jenkins.sh, for gerrit build bot

  [ pierre.baudry ]
  * transceiver: Fix mismatched allocations and deallocations

  [ Holger Hans Peter Freyther ]
  * debian: Require fftw3 header files for osmo-trx

  [ Max ]
  * Add gerrit settings
  * Integrate Debian packaging changes
  * Remove embedded sqlite3
  * Fix building against sqlite3
  * Add autoconf-archive to dependencies
  * debian: remove obsolete dependency
  * deb: remove unused dependency
  * Remove redundant explicit dependency
  * Use release helper from libosmocore

  [ Ruben Undheim ]
  * Do not embed sqlite3 when building

  [ Philipp Maier ]
  * buildenv: Turn off native architecture builds
  * cosmetic: Make parameter lists uniform
  * Add test program to verify convolution implementation
  * ssedetect: Add runtime CPU detection
  * cosmetic: remove code duplication
  * buildenv: Make build CPU invariant
  * buildenv: Split up SSE3 and SSE4.1 code
  * cosmetic: Add info about SSE support

  [ Vadim Yanitskiy ]
  * buildenv: correct the ax_sse macro description
  * buildenv: actually strip unused cpuid functionality
  * buildenv: fix build on systems without SIMD support
  * buildenv: cosmetic changes
  * buildenv: check for __builtin_cpu_supports call support
  * ssedetect: call __builtin_cpu_supports() only if supported

  [ Pau Espin Pedrol ]
  * cosmetic: transciever: Remove trailing whitespaces
  * transceiver: Avoid sending clock indications when trx is not powered on
  * Add -j option to bind to specific address

  [ ignasj ]
  * LimeSDR: Change device detection to work with USB and PCIe versions
  * LimeSDR: change tx window type to TX_WINDOW_FIXED
  * LimeSDR: Fix sample value range

  [ Harald Welte ]
  * Add '-t' command line option to enable SCHED_RR
  * Import git-version-gen and update AC_INIT()

Change-Id: Ibf3be6cc25e9b20d625b1f67972114b7f613f05c

.gitignore

@@ -3,8 +3,6 @@
 *.lo
 *.la
 Transceiver52M/osmo-trx
-Transceiver52M/osmo-trx-gen
-Transceiver52M/osmo-trx-dec
 
 # tests
 CommonLibs/BitVectorTest

CommonLibs/Sockets.cpp

@@ -223,18 +223,18 @@ int DatagramSocket::read(char* buffer, size_t length, unsigned timeout)
 
 
 
-UDPSocket::UDPSocket(unsigned short wSrcPort)
+UDPSocket::UDPSocket(const char *wSrcIP, unsigned short wSrcPort)
 	:DatagramSocket()
 {
-	open(wSrcPort);
+	open(wSrcPort, wSrcIP);
 }
 
 
-UDPSocket::UDPSocket(unsigned short wSrcPort,
+UDPSocket::UDPSocket(const char *wSrcIP, unsigned short wSrcPort,
 		     const char *wDestIP, unsigned short wDestPort)
 	:DatagramSocket()
 {
-	open(wSrcPort);
+	open(wSrcPort, wSrcIP);
 	destination(wDestPort, wDestIP);
 }
 
@@ -246,7 +246,7 @@ void UDPSocket::destination( unsigned short wDestPort, const char * wDestIP )
 }
 
 
-void UDPSocket::open(unsigned short localPort)
+void UDPSocket::open(unsigned short localPort, const char *wlocalIP)
 {
 	// create
 	mSocketFD = socket(AF_INET,SOCK_DGRAM,0);
@@ -265,7 +265,7 @@ void UDPSocket::open(unsigned short localPort)
 	size_t length = sizeof(address);
 	bzero(&address,length);
 	address.sin_family = AF_INET;
-	address.sin_addr.s_addr = htonl(INADDR_LOOPBACK);
+	address.sin_addr.s_addr = inet_addr(wlocalIP);
 	address.sin_port = htons(localPort);
 	if (bind(mSocketFD,(struct sockaddr*)&address,length)<0) {
 		perror("bind() failed");

CommonLibs/Sockets.h

@@ -144,10 +144,10 @@ class UDPSocket : public DatagramSocket {
 public:
 
 	/** Open a USP socket with an OS-assigned port and no default destination. */
-	UDPSocket( unsigned short localPort=0);
+	UDPSocket(const char *localIP, unsigned short localPort);
 
 	/** Given a full specification, open the socket and set the dest address. */
-	UDPSocket( 	unsigned short localPort, 
+	UDPSocket(const char *localIP, unsigned short localPort,
 		  const char *remoteIP, unsigned short remotePort);
 
 	/** Set the destination port. */
@@ -157,7 +157,7 @@ public:
 	unsigned short port() const;
 
 	/** Open and bind the UDP socket to a local port. */
-	void open(unsigned short localPort=0);
+	void open(unsigned short localPort=0, const char *wlocalIP="127.0.0.1");
 
 	/** Give the return address of the most recently received packet. */
 	const struct sockaddr_in* source() const { return (const struct sockaddr_in*)mSource; }

CommonLibs/SocketsTest.cpp

@@ -37,7 +37,7 @@ static const int gNumToSend = 10;
 
 void *testReaderIP(void *)
 {
-	UDPSocket readSocket(5934, "localhost", 5061);
+	UDPSocket readSocket("127.0.0.1", 5934, "localhost", 5061);
 	readSocket.nonblocking();
 	int rc = 0;
 	while (rc<gNumToSend) {
@@ -82,7 +82,7 @@ int main(int argc, char * argv[] )
   Thread readerThreadUnix;
   readerThreadUnix.start(testReaderUnix,NULL);
 
-  UDPSocket socket1(5061, "127.0.0.1",5934);
+  UDPSocket socket1("127.0.0.1", 5061, "127.0.0.1", 5934);
   UDDSocket socket1U("testSource","testDestination");
   
   COUT("socket1: " << socket1.port());

CommonLibs/Vector.h

@@ -118,8 +118,8 @@ template <class T> class Vector {
 	/** Build an empty Vector of a given size. */
 	Vector(size_t wSize=0):mData(NULL) { resize(wSize); }
 
-	/** Build a Vector by shifting the data block. */
+	/** Build a Vector by moving another. */
-	Vector(Vector<T>& other)
+	Vector(Vector<T>&& other)
 		:mData(other.mData),mStart(other.mStart),mEnd(other.mEnd)
 	{ other.mData=NULL; }
 

Makefile.am

@@ -39,6 +39,7 @@ EXTRA_DIST = \
 	COPYING \
 	README
 
+@RELMAKE@
 
 dox: FORCE
 	doxygen doxconfig

Transceiver52M/Makefile.am

@@ -69,10 +69,7 @@ libtransceiver_la_SOURCES = \
 	radioInterfaceResamp.cpp \
 	radioInterfaceMulti.cpp
 
-bin_PROGRAMS = \
+bin_PROGRAMS = osmo-trx
-	osmo-trx \
-	osmo-trx-gen \
-	osmo-trx-dec
 
 noinst_HEADERS = \
 	Complex.h \
@@ -102,20 +99,6 @@ osmo_trx_LDADD = \
 	$(GSM_LA) \
 	$(COMMON_LA) $(SQLITE3_LIBS)
 
-osmo_trx_gen_SOURCES = osmo-trx-gen.cpp
-osmo_trx_gen_LDADD = \
-	libtransceiver.la \
-	$(ARCH_LA) \
-	$(GSM_LA) \
-	$(COMMON_LA) $(SQLITE_LA)
-
-osmo_trx_dec_SOURCES = osmo-trx-dec.cpp
-osmo_trx_dec_LDADD = \
-	libtransceiver.la \
-	$(ARCH_LA) \
-	$(GSM_LA) \
-	$(COMMON_LA) $(SQLITE_LA)
-
 if USRP1 
 libtransceiver_la_SOURCES += USRPDevice.cpp
 osmo_trx_LDADD += $(USRP_LIBS)

Transceiver52M/Resampler.cpp

@@ -22,6 +22,7 @@
 #include <string.h>
 #include <malloc.h>
 #include <iostream>
+#include <algorithm>
 
 #include "Resampler.h"
 
@@ -35,6 +36,8 @@ extern "C" {
 
 #define MAX_OUTPUT_LEN		4096
 
+using namespace std;
+
 static float sinc(float x)
 {
 	if (x == 0.0)
@@ -43,32 +46,19 @@ static float sinc(float x)
 	return sin(M_PI * x) / (M_PI * x);
 }
 
-bool Resampler::initFilters(float bw)
+void Resampler::initFilters(float bw)
 {
-	size_t proto_len = p * filt_len;
+	float cutoff;
-	float *proto, val, cutoff;
 	float sum = 0.0f, scale = 0.0f;
-	float midpt = (float) (proto_len - 1.0) / 2.0;
 
 	/* 
 	 * Allocate partition filters and the temporary prototype filter
 	 * according to numerator of the rational rate. Coefficients are
 	 * real only and must be 16-byte memory aligned for SSE usage.
 	 */
-	proto = new float[proto_len];
+	auto proto = vector<float>(p * filt_len);
-	if (!proto)
+	for (auto &part : partitions)
-		return false;
+		part = (complex<float> *) memalign(16, filt_len * sizeof(complex<float>));
-
-	partitions = (float **) malloc(sizeof(float *) * p);
-	if (!partitions) {
-		delete[] proto;
-		return false;
-	}
-
-	for (size_t i = 0; i < p; i++) {
-		partitions[i] = (float *)
-				memalign(16, filt_len * 2 * sizeof(float));
-	}
 
 	/* 
 	 * Generate the prototype filter with a Blackman-harris window.
@@ -85,47 +75,26 @@ bool Resampler::initFilters(float bw)
 	else
 		cutoff = (float) q;
 
-	for (size_t i = 0; i < proto_len; i++) {
+	float midpt = (proto.size() - 1) / 2.0;
+	for (size_t i = 0; i < proto.size(); i++) {
 		proto[i] = sinc(((float) i - midpt) / cutoff * bw);
 		proto[i] *= a0 -
-			    a1 * cos(2 * M_PI * i / (proto_len - 1)) +
+			    a1 * cos(2 * M_PI * i / (proto.size() - 1)) +
-			    a2 * cos(4 * M_PI * i / (proto_len - 1)) -
+			    a2 * cos(4 * M_PI * i / (proto.size() - 1)) -
-			    a3 * cos(6 * M_PI * i / (proto_len - 1));
+			    a3 * cos(6 * M_PI * i / (proto.size() - 1));
 		sum += proto[i];
 	}
 	scale = p / sum;
 
 	/* Populate filter partitions from the prototype filter */
 	for (size_t i = 0; i < filt_len; i++) {
-		for (size_t n = 0; n < p; n++) {
+		for (size_t n = 0; n < p; n++)
-			partitions[n][2 * i + 0] = proto[i * p + n] * scale;
+			partitions[n][i] = complex<float>(proto[i * p + n] * scale);
-			partitions[n][2 * i + 1] = 0.0f;
-		}
 	}
 
-	/* For convolution, we store the filter taps in reverse */ 
+	/* Store filter taps in reverse */
-	for (size_t n = 0; n < p; n++) {
+	for (auto &part : partitions)
-		for (size_t i = 0; i < filt_len / 2; i++) {
+		reverse(&part[0], &part[filt_len]);
-			val = partitions[n][2 * i];
-			partitions[n][2 * i] = partitions[n][2 * (filt_len - 1 - i)];
-			partitions[n][2 * (filt_len - 1 - i)] = val;
-		}
-	}
-
-	delete[] proto;
-
-	return true;
-}
-
-void Resampler::releaseFilters()
-{
-	if (partitions) {
-		for (size_t i = 0; i < p; i++)
-			free(partitions[i]);
-	}
-
-	free(partitions);
-	partitions = NULL;
 }
 
 static bool check_vec_len(int in_len, int out_len, int p, int q)
@@ -159,14 +128,6 @@ static bool check_vec_len(int in_len, int out_len, int p, int q)
 	return true;
 }
 
-void Resampler::computePath()
-{
-	for (int i = 0; i < MAX_OUTPUT_LEN; i++) {
-		in_index[i] = (q * i) / p;
-		out_path[i] = (q * i) % p;
-	}
-}
-
 int Resampler::rotate(const float *in, size_t in_len, float *out, size_t out_len)
 {
 	int n, path;
@@ -180,8 +141,8 @@ int Resampler::rotate(const float *in, size_t in_len, float *out, size_t out_len
 		path = out_path[i]; 
 
 		convolve_real(in, in_len,
-			      partitions[path], filt_len,
+			      reinterpret_cast<float *>(partitions[path]),
-			      &out[2 * i], out_len - i,
+			      filt_len, &out[2 * i], out_len - i,
 			      n, 1, 1, 0);
 	}
 
@@ -190,14 +151,18 @@ int Resampler::rotate(const float *in, size_t in_len, float *out, size_t out_len
 
 bool Resampler::init(float bw)
 {
+	if (p == 0 || q == 0 || filt_len == 0) return false;
+
 	/* Filterbank filter internals */
-	if (!initFilters(bw))
+	initFilters(bw);
-		return false;
 
 	/* Precompute filterbank paths */
-	in_index = new size_t[MAX_OUTPUT_LEN];
+	int i = 0;
-	out_path = new size_t[MAX_OUTPUT_LEN];
+	for (auto &index : in_index)
-	computePath();
+		index = (q * i++) / p;
+	i = 0;
+	for (auto &path : out_path)
+		path = (q * i++) % p;
 
 	return true;
 }
@@ -208,7 +173,7 @@ size_t Resampler::len()
 }
 
 Resampler::Resampler(size_t p, size_t q, size_t filt_len)
-	: in_index(NULL), out_path(NULL), partitions(NULL)
+	: in_index(MAX_OUTPUT_LEN), out_path(MAX_OUTPUT_LEN), partitions(p)
 {
 	this->p = p;
 	this->q = q;
@@ -217,8 +182,6 @@ Resampler::Resampler(size_t p, size_t q, size_t filt_len)
 
 Resampler::~Resampler()
 {
-	releaseFilters();
+	for (auto &part : partitions)
-
+		free(part);
-	delete in_index;
-	delete out_path;
 }

Transceiver52M/Resampler.h

@@ -20,6 +20,9 @@
 #ifndef _RESAMPLER_H_
 #define _RESAMPLER_H_
 
+#include <vector>
+#include <complex>
+
 class Resampler {
 public:
 	/* Constructor for rational sample rate conversion
@@ -63,14 +66,11 @@ private:
 	size_t p;
 	size_t q;
 	size_t filt_len;
-	size_t *in_index;
+	std::vector<size_t> in_index;
-	size_t *out_path;
+	std::vector<size_t> out_path;
+	std::vector<std::complex<float> *> partitions;
 
-	float **partitions;
+	void initFilters(float bw);
-
-	bool initFilters(float bw);
-	void releaseFilters();
-	void computePath();
 };
 
 #endif /* _RESAMPLER_H_ */

Transceiver52M/Transceiver.cpp

@@ -85,7 +85,7 @@ bool TransceiverState::init(int filler, size_t sps, float scale, size_t rtsc, un
         burst = generateDummyBurst(sps, n);
         break;
       case Transceiver::FILLER_NORM_RAND:
-        burst = genRandNormalBurst(rtsc, sps, n, mPrbs);
+        burst = genRandNormalBurst(rtsc, sps, n);
         break;
       case Transceiver::FILLER_EDGE_RAND:
         burst = generateEdgeBurst(rtsc);
@@ -112,16 +112,17 @@ bool TransceiverState::init(int filler, size_t sps, float scale, size_t rtsc, un
 }
 
 Transceiver::Transceiver(int wBasePort,
-                         const char *wTRXAddress,
+                         const char *TRXAddress,
+                         const char *GSMcoreAddress,
                          size_t tx_sps, size_t rx_sps, size_t chans,
                          GSM::Time wTransmitLatency,
                          RadioInterface *wRadioInterface,
                          double wRssiOffset)
-  : mBasePort(wBasePort), mAddr(wTRXAddress),
+  : mBasePort(wBasePort), mLocalAddr(TRXAddress), mRemoteAddr(GSMcoreAddress),
-    mClockSocket(wBasePort, wTRXAddress, mBasePort + 100),
+    mClockSocket(TRXAddress, wBasePort, GSMcoreAddress, wBasePort + 100),
     mTransmitLatency(wTransmitLatency), mRadioInterface(wRadioInterface),
     rssiOffset(wRssiOffset),
-    mSPSTx(tx_sps), mSPSRx(rx_sps), mChans(chans), mEdge(false), mOn(false),
+    mSPSTx(tx_sps), mSPSRx(rx_sps), mChans(chans), mEdge(false), mOn(false), mForceClockInterface(false),
     mTxFreq(0.0), mRxFreq(0.0), mTSC(0), mMaxExpectedDelayAB(0), mMaxExpectedDelayNB(0),
     mWriteBurstToDiskMask(0)
 {
@@ -197,8 +198,8 @@ bool Transceiver::init(int filler, size_t rtsc, unsigned rach_delay, bool edge)
     d_srcport = mBasePort + 2 * i + 2;
     d_dstport = mBasePort + 2 * i + 102;
 
-    mCtrlSockets[i] = new UDPSocket(c_srcport, mAddr.c_str(), c_dstport);
+    mCtrlSockets[i] = new UDPSocket(mLocalAddr.c_str(), c_srcport, mRemoteAddr.c_str(), c_dstport);
-    mDataSockets[i] = new UDPSocket(d_srcport, mAddr.c_str(), d_dstport);
+    mDataSockets[i] = new UDPSocket(mLocalAddr.c_str(), d_srcport, mRemoteAddr.c_str(), d_dstport);
   }
 
   /* Randomize the central clock */
@@ -273,7 +274,7 @@ bool Transceiver::start()
                             TxUpperLoopAdapter, (void*) chan);
   }
 
-  writeClockInterface();
+  mForceClockInterface = true;
   mOn = true;
   return true;
 }
@@ -297,6 +298,10 @@ void Transceiver::stop()
   LOG(NOTICE) << "Stopping the transceiver";
   mTxLowerLoopThread->cancel();
   mRxLowerLoopThread->cancel();
+  mTxLowerLoopThread->join();
+  mRxLowerLoopThread->join();
+  delete mTxLowerLoopThread;
+  delete mRxLowerLoopThread;
 
   for (size_t i = 0; i < mChans; i++) {
     mRxServiceLoopThreads[i]->cancel();
@@ -315,11 +320,6 @@ void Transceiver::stop()
     mTxPriorityQueues[i].clear();
   }
 
-  mTxLowerLoopThread->join();
-  mRxLowerLoopThread->join();
-  delete mTxLowerLoopThread;
-  delete mRxLowerLoopThread;
-
   mOn = false;
   LOG(NOTICE) << "Transceiver stopped";
 }
@@ -607,7 +607,7 @@ SoftVector *Transceiver::pullRadioVector(GSM::Time &wTime, double &RSSI, bool &i
   avg = sqrt(avg / radio_burst->chans());
 
   wTime = time;
-  RSSI = 20.0 * log10(avg / rxFullScale);
+  RSSI = 20.0 * log10(rxFullScale / avg);
 
   /* RSSI estimation are valid */
   isRssiValid = true;
@@ -678,9 +678,6 @@ void Transceiver::driveControl(size_t chan)
 
   sscanf(buffer,"%3s %s",cmdcheck,command);
 
-  if (!chan)
-    writeClockInterface();
-
   if (strcmp(cmdcheck,"CMD")!=0) {
     LOG(WARNING) << "bogus message on control interface";
     return;
@@ -874,8 +871,8 @@ void Transceiver::driveReceiveRadio()
 {
   if (!mRadioInterface->driveReceiveRadio()) {
     usleep(100000);
-  } else {
+  } else if (mForceClockInterface || mTransmitDeadlineClock > mLastClockUpdateTime + GSM::Time(216,0)) {
-    if (mTransmitDeadlineClock > mLastClockUpdateTime + GSM::Time(216,0))
+    mForceClockInterface = false;
     writeClockInterface();
   }
 }

Transceiver52M/Transceiver.h

@@ -84,9 +84,6 @@ struct TransceiverState {
 
   /* Shadowed downlink attenuation */
   int mPower;
-
-  /* Pseudorandom bit sequence */
-  PRBS9 mPrbs;
 };
 
 /** The Transceiver class, responsible for physical layer of basestation */
@@ -94,13 +91,15 @@ class Transceiver {
 public:
   /** Transceiver constructor
       @param wBasePort base port number of UDP sockets
-      @param TRXAddress IP address of the TRX manager, as a string
+      @param TRXAddress IP address of the TRX, as a string
+      @param GSMcoreAddress IP address of the GSM core, as a string
       @param wSPS number of samples per GSM symbol
       @param wTransmitLatency initial setting of transmit latency
       @param radioInterface associated radioInterface object
   */
   Transceiver(int wBasePort,
               const char *TRXAddress,
+              const char *GSMcoreAddress,
               size_t tx_sps, size_t rx_sps, size_t chans,
               GSM::Time wTransmitLatency,
               RadioInterface *wRadioInterface,
@@ -155,7 +154,8 @@ public:
 
 private:
   int mBasePort;
-  std::string mAddr;
+  std::string mLocalAddr;
+  std::string mRemoteAddr;
 
   std::vector<UDPSocket *> mDataSockets;  ///< socket for writing to/reading from GSM core
   std::vector<UDPSocket *> mCtrlSockets;  ///< socket for writing/reading control commands from GSM core
@@ -211,6 +211,7 @@ private:
 
   bool mEdge;
   bool mOn;	                           ///< flag to indicate that transceiver is powered on
+  bool mForceClockInterface;           ///< flag to indicate whether IND CLOCK shall be sent unconditionally after transceiver is started
   bool mHandover[8][8];                ///< expect handover to the timeslot/subslot
   double mTxFreq;                      ///< the transmit frequency
   double mRxFreq;                      ///< the receive frequency
@@ -278,4 +279,3 @@ void *ControlServiceLoopAdapter(TransceiverChannel *);
 
 /** transmit queueing thread loop */
 void *TxUpperLoopAdapter(TransceiverChannel *);
-

Transceiver52M/UHDDevice.cpp

@@ -21,6 +21,7 @@
  * See the COPYING file in the main directory for details.
  */
 
+#include <map>
 #include "radioDevice.h"
 #include "Threads.h"
 #include "Logger.h"
@@ -37,12 +38,6 @@
 #include <uhd/utils/msg.hpp>
 #endif
 
-#define B2XX_CLK_RT      26e6
-#define B2XX_MCBTS_CLK_RT   51.2e6
-#define E1XX_CLK_RT      52e6
-#define LIMESDR_CLK_RT   (GSMRATE*32)
-#define B100_BASE_RT     400000
-#define USRP2_BASE_RT    390625
 #define USRP_TX_AMPL     0.3
 #define UMTRX_TX_AMPL    0.7
 #define LIMESDR_TX_AMPL  0.3
@@ -73,15 +68,6 @@ enum uhd_dev_type {
 	X3XX,
 	UMTRX,
 	LIMESDR,
-	NUM_USRP_TYPES,
-};
-
-struct uhd_dev_offset {
-	enum uhd_dev_type type;
-	size_t tx_sps;
-	size_t rx_sps;
-	double offset;
-	const std::string desc;
 };
 
 /*
@@ -109,76 +95,44 @@ struct uhd_dev_offset {
  * Notes:
  *   USRP1 with timestamps is not supported by UHD.
  */
-static struct uhd_dev_offset uhd_offsets[] = {
+
-	{ USRP1, 1, 1,       0.0, "USRP1 not supported" },
+/* Device Type, Tx-SPS, Rx-SPS */
-	{ USRP1, 4, 1,       0.0, "USRP1 not supported"},
+typedef std::tuple<uhd_dev_type, int, int> dev_key;
-	{ USRP2, 1, 1, 1.2184e-4, "N2XX 1 SPS" },
+
-	{ USRP2, 4, 1, 7.6547e-5, "N2XX 4/1 SPS" },
+/* Device parameter descriptor */
-	{ B100,  1, 1, 1.2104e-4, "B100 1 SPS" },
+struct dev_desc {
-	{ B100,  4, 1, 7.9307e-5, "B100 4 SPS" },
+	unsigned channels;
-	{ B200,  1, 1, B2XX_TIMING_1SPS, "B200 1 SPS" },
+	double mcr;
-	{ B200,  4, 1, B2XX_TIMING_4SPS, "B200 4/1 Tx/Rx SPS" },
+	double rate;
-	{ B210,  1, 1, B2XX_TIMING_1SPS, "B210 1 SPS" },
+	double offset;
-	{ B210,  4, 1, B2XX_TIMING_4SPS, "B210 4/1 Tx/Rx SPS" },
+	std::string str;
-	{ B2XX_MCBTS, 4, 4, B2XX_TIMING_MCBTS, "B200/B210 4 SPS Multi-ARFCN" },
-	{ E1XX,  1, 1, 9.5192e-5, "E1XX 1 SPS" },
-	{ E1XX,  4, 1, 6.5571e-5, "E1XX 4/1 Tx/Rx SPS" },
-	{ E3XX,  1, 1, 1.84616e-4, "E3XX 1 SPS" },
-	{ E3XX,  4, 1, 1.29231e-4, "E3XX 4/1 Tx/Rx SPS" },
-	{ X3XX,  1, 1, 1.5360e-4, "X3XX 1 SPS"},
-	{ X3XX,  4, 1, 1.1264e-4, "X3XX 4/1 Tx/Rx SPS"},
-	{ UMTRX, 1, 1, 9.9692e-5, "UmTRX 1 SPS" },
-	{ UMTRX, 4, 1, 7.3846e-5, "UmTRX 4/1 Tx/Rx SPS" },
-	{ USRP2, 4, 4, 4.6080e-5, "N2XX 4 SPS" },
-	{ B200,  4, 4, B2XX_TIMING_4_4SPS, "B200 4 SPS" },
-	{ B210,  4, 4, B2XX_TIMING_4_4SPS, "B210 4 SPS" },
-	{ X3XX,  4, 4, 5.6567e-5, "X3XX 4 SPS"},
-	{ UMTRX, 4, 4, 5.1503e-5, "UmTRX 4 SPS" },
-	{ LIMESDR, 4, 4, 16.5/GSMRATE, "STREAM/LimeSDR (4 SPS TX/RX)" },
 };
-#define NUM_UHD_OFFSETS (sizeof(uhd_offsets)/sizeof(uhd_offsets[0]))
 
-/*
+static const std::map<dev_key, dev_desc> dev_param_map {
- * Select sample rate based on device type and requested samples-per-symbol.
+	{ std::make_tuple(USRP2, 1, 1), { 1, 0.0,  390625,  1.2184e-4,  "N2XX 1 SPS"         } },
- * The base rate is either GSM symbol rate, 270.833 kHz, or the minimum
+	{ std::make_tuple(USRP2, 4, 1), { 1, 0.0,  390625,  7.6547e-5,  "N2XX 4/1 Tx/Rx SPS" } },
- * usable channel spacing of 400 kHz.
+	{ std::make_tuple(USRP2, 4, 4), { 1, 0.0,  390625,  4.6080e-5,  "N2XX 4 SPS"         } },
- */
+	{ std::make_tuple(B100,  1, 1), { 1, 0.0,  400000,  1.2104e-4,  "B100 1 SPS"         } },
-static double select_rate(uhd_dev_type type, int sps,
+	{ std::make_tuple(B100,  4, 1), { 1, 0.0,  400000,  7.9307e-5,  "B100 4/1 Tx/Rx SPS" } },
-			  RadioDevice::InterfaceType iface)
+	{ std::make_tuple(B200,  1, 1), { 1, 26e6, GSMRATE, B2XX_TIMING_1SPS, "B200 1 SPS"   } },
-{
+	{ std::make_tuple(B200,  4, 1), { 1, 26e6, GSMRATE, B2XX_TIMING_4SPS, "B200 4/1 Tx/Rx SPS" } },
-	if ((sps != 4) && (sps != 1))
+	{ std::make_tuple(B200,  4, 4), { 1, 26e6, GSMRATE, B2XX_TIMING_4_4SPS, "B200 4 SPS" } },
-		return -9999.99;
+	{ std::make_tuple(B210,  1, 1), { 2, 26e6, GSMRATE, B2XX_TIMING_1SPS, "B210 1 SPS"    } },
-
+	{ std::make_tuple(B210,  4, 1), { 2, 26e6, GSMRATE, B2XX_TIMING_4SPS, "B210 4/1 Tx/Rx SPS" } },
-	if (iface == RadioDevice::MULTI_ARFCN) {
+	{ std::make_tuple(B210,  4, 4), { 2, 26e6, GSMRATE, B2XX_TIMING_4_4SPS, "B210 4 SPS" } },
-		switch (type) {
+	{ std::make_tuple(E1XX,  1, 1), { 1, 52e6, GSMRATE, 9.5192e-5,  "E1XX 1 SPS"         } },
-		case B2XX_MCBTS:
+	{ std::make_tuple(E1XX,  4, 1), { 1, 52e6, GSMRATE, 6.5571e-5,  "E1XX 4/1 Tx/Rx SPS" } },
-			return  4 * MCBTS_SPACING;
+	{ std::make_tuple(E3XX,  1, 1), { 2, 26e6, GSMRATE, 1.8462e-4,  "E3XX 1 SPS"         } },
-		default:
+	{ std::make_tuple(E3XX,  4, 1), { 2, 26e6, GSMRATE, 1.2923e-4,  "E3XX 4/1 Tx/Rx SPS" } },
-			LOG(ALERT) << "Invalid device combination";
+	{ std::make_tuple(X3XX,  1, 1), { 2, 0.0,  390625,  1.5360e-4,  "X3XX 1 SPS"         } },
-			return -9999.99;
+	{ std::make_tuple(X3XX,  4, 1), { 2, 0.0,  390625,  1.1264e-4,  "X3XX 4/1 Tx/Rx SPS" } },
-		}
+	{ std::make_tuple(X3XX,  4, 4), { 2, 0.0,  390625,  5.6567e-5,  "X3XX 4 SPS"         } },
-	}
+	{ std::make_tuple(UMTRX, 1, 1), { 2, 0.0,  GSMRATE, 9.9692e-5,  "UmTRX 1 SPS"        } },
-
+	{ std::make_tuple(UMTRX, 4, 1), { 2, 0.0,  GSMRATE, 7.3846e-5,  "UmTRX 4/1 Tx/Rx SPS"} },
-	switch (type) {
+	{ std::make_tuple(UMTRX, 4, 4), { 2, 0.0,  GSMRATE, 5.1503e-5,  "UmTRX 4 SPS"        } },
-	case USRP2:
+	{ std::make_tuple(LIMESDR, 4, 4), { 1, GSMRATE*32, GSMRATE, 8.9e-5, "LimeSDR 4 SPS"  } },
-	case X3XX:
+	{ std::make_tuple(B2XX_MCBTS, 4, 4), { 1, 51.2e6, MCBTS_SPACING*4, B2XX_TIMING_MCBTS, "B200/B210 4 SPS Multi-ARFCN" } },
-		return USRP2_BASE_RT * sps;
+};
-	case B100:
-		return B100_BASE_RT * sps;
-	case B200:
-	case B210:
-	case E1XX:
-	case E3XX:
-	case UMTRX:
-	case LIMESDR:
-		return GSMRATE * sps;
-	default:
-		break;
-	}
-
-	LOG(ALERT) << "Unknown device type " << type;
-	return -9999.99;
-}
 
 /*
     Sample Buffer - Allows reading and writing of timed samples using osmo-trx
@@ -339,9 +293,8 @@ private:
 	std::vector<smpl_buf *> rx_buffers;
 
 	void init_gains();
-	double get_dev_offset();
+	void set_channels(bool swap);
-	int set_master_clk(double rate);
+	void set_rates();
-	int set_rates(double tx_rate, double rx_rate);
 	bool parse_dev_type();
 	bool flush_recv(size_t num_pkts);
 	int check_rx_md_err(uhd::rx_metadata_t &md, ssize_t num_smpls);
@@ -470,105 +423,22 @@ void uhd_device::init_gains()
 
 }
 
-double uhd_device::get_dev_offset()
+void uhd_device::set_rates()
 {
-	struct uhd_dev_offset *offset = NULL;
+	dev_desc desc = dev_param_map.at(dev_key(dev_type, tx_sps, rx_sps));
+	if (desc.mcr != 0.0)
+		usrp_dev->set_master_clock_rate(desc.mcr);
 
-	/* Reject USRP1 */
+	tx_rate = (dev_type != B2XX_MCBTS) ? desc.rate * tx_sps : desc.rate;
-	if (dev_type == USRP1) {
+	rx_rate = (dev_type != B2XX_MCBTS) ? desc.rate * rx_sps : desc.rate;
-		LOG(ERR) << "Invalid device type";
-		return 0.0;
-	}
 
-	/* Search for matching offset value */
-	for (size_t i = 0; i < NUM_UHD_OFFSETS; i++) {
-		if ((dev_type == uhd_offsets[i].type) &&
-			(tx_sps == uhd_offsets[i].tx_sps) &&
-			(rx_sps == uhd_offsets[i].rx_sps)) {
-			offset = &uhd_offsets[i];
-			break;
-		}
-	}
-
-	if (!offset) {
-		LOG(ERR) << "Invalid device configuration";
-		return 0.0;
-	}
-
-	std::cout << "-- Setting " << offset->desc << std::endl;
-
-	return offset->offset;
-}
-
-int uhd_device::set_master_clk(double clk_rate)
-{
-	double actual, offset, limit = 1.0;
-
-	try {
-		usrp_dev->set_master_clock_rate(clk_rate);
-	} catch (const std::exception &ex) {
-		LOG(ALERT) << "UHD clock rate setting failed: " << clk_rate;
-		LOG(ALERT) << ex.what();
-		return -1;
-	}
-
-	actual = usrp_dev->get_master_clock_rate();
-	offset = fabs(clk_rate - actual);
-
-	if (offset > limit) {
-		LOG(ALERT) << "Failed to set master clock rate";
-		LOG(ALERT) << "Requested clock rate " << clk_rate;
-		LOG(ALERT) << "Actual clock rate " << actual;
-		return -1;
-	}
-
-	return 0;
-}
-
-int uhd_device::set_rates(double tx_rate, double rx_rate)
-{
-	double offset_limit = 1.0;
-	double tx_offset, rx_offset;
-
-	/* B2XX and E1xx are the only device where we set FPGA clocking */
-	if ((dev_type == B200) || (dev_type == B210) || (dev_type == E3XX)) {
-		if (set_master_clk(B2XX_CLK_RT) < 0)
-			return -1;
-	} else if (dev_type == E1XX) {
-		if (set_master_clk(E1XX_CLK_RT) < 0)
-			return -1;
-	} else if (dev_type == B2XX_MCBTS) {
-		if (set_master_clk(B2XX_MCBTS_CLK_RT) < 0)
-			return -1;
-	}
-	else if (dev_type == LIMESDR) {
-		if (set_master_clk(LIMESDR_CLK_RT) < 0)
-			return -1;
-	}
-
-
-	// Set sample rates
-	try {
 	usrp_dev->set_tx_rate(tx_rate);
 	usrp_dev->set_rx_rate(rx_rate);
-	} catch (const std::exception &ex) {
+	tx_rate = usrp_dev->get_tx_rate();
-		LOG(ALERT) << "UHD rate setting failed";
+	rx_rate = usrp_dev->get_rx_rate();
-		LOG(ALERT) << ex.what();
-		return -1;
-	}
-	this->tx_rate = usrp_dev->get_tx_rate();
-	this->rx_rate = usrp_dev->get_rx_rate();
 
-	tx_offset = fabs(this->tx_rate - tx_rate);
+	ts_offset = static_cast<TIMESTAMP>(desc.offset * rx_rate);
-	rx_offset = fabs(this->rx_rate - rx_rate);
+	LOG(INFO) << "Rates configured for " << desc.str;
-	if ((tx_offset > offset_limit) || (rx_offset > offset_limit)) {
-		LOG(ALERT) << "Actual sample rate differs from desired rate";
-		LOG(ALERT) << "Tx/Rx (" << this->tx_rate << "/"
-			   << this->rx_rate << ")";
-		return -1;
-	}
-
-	return 0;
 }
 
 double uhd_device::setTxGain(double db, size_t chan)
@@ -641,87 +511,41 @@ double uhd_device::getRxGain(size_t chan)
  */
 bool uhd_device::parse_dev_type()
 {
-	std::string mboard_str, dev_str;
+	uhd::property_tree::sptr prop_tree = usrp_dev->get_device()->get_tree();
-	uhd::property_tree::sptr prop_tree;
+	std::string devString = prop_tree->access<std::string>("/name").get();
-	size_t usrp1_str, usrp2_str, e100_str, e110_str, e310_str, e3xx_str,
+	std::string mboardString = usrp_dev->get_mboard_name();
-	       b100_str, b200_str, b210_str, x300_str, x310_str, umtrx_str, limesdr_str;
 
-	prop_tree = usrp_dev->get_device()->get_tree();
+	const std::map<std::string, std::pair<uhd_dev_type, TxWindowType>> devStringMap {
-	dev_str = prop_tree->access<std::string>("/name").get();
+		{ "B100",     { B100,    TX_WINDOW_USRP1 } },
-	mboard_str = usrp_dev->get_mboard_name();
+		{ "B200",     { B200,    TX_WINDOW_USRP1 } },
-
+		{ "B200mini", { B200,    TX_WINDOW_USRP1 } },
-	usrp1_str = dev_str.find("USRP1");
+		{ "B210",     { B210,    TX_WINDOW_USRP1 } },
-	usrp2_str = dev_str.find("USRP2");
+		{ "E100",     { E1XX,    TX_WINDOW_FIXED } },
-	b100_str = mboard_str.find("B100");
+		{ "E110",     { E1XX,    TX_WINDOW_FIXED } },
-	b200_str = mboard_str.find("B200");
+		{ "E310",     { E3XX,    TX_WINDOW_FIXED } },
-	b210_str = mboard_str.find("B210");
+		{ "E3XX",     { E3XX,    TX_WINDOW_FIXED } },
-	e100_str = mboard_str.find("E100");
+		{ "X300",     { X3XX,    TX_WINDOW_FIXED } },
-	e110_str = mboard_str.find("E110");
+		{ "X310",     { X3XX,    TX_WINDOW_FIXED } },
-	e310_str = mboard_str.find("E310");
+		{ "USRP2",    { USRP2,   TX_WINDOW_FIXED } },
-	e3xx_str = mboard_str.find("E3XX");
+		{ "UmTRX",    { UMTRX,   TX_WINDOW_FIXED } },
-	x300_str = mboard_str.find("X300");
+		{ "LimeSDR",  { LIMESDR, TX_WINDOW_FIXED } },
-	x310_str = mboard_str.find("X310");
+	};
-	umtrx_str = dev_str.find("UmTRX");
-	// LimeSDR is based on STREAM board, so it's advertized as such
-	limesdr_str = dev_str.find("STREAM");
-
-	if (usrp1_str != std::string::npos) {
-		LOG(ALERT) << "USRP1 is not supported using the UHD driver";
-		LOG(ALERT) << "Please compile with GNU Radio libusrp support";
-		dev_type = USRP1;
-		return false;
-	}
-
-	if (b100_str != std::string::npos) {
-		tx_window = TX_WINDOW_USRP1;
-		dev_type = B100;
-	} else if (b200_str != std::string::npos) {
-		tx_window = TX_WINDOW_USRP1;
-		dev_type = B200;
-	} else if (b210_str != std::string::npos) {
-		tx_window = TX_WINDOW_USRP1;
-		dev_type = B210;
-	} else if (e100_str != std::string::npos) {
-		tx_window = TX_WINDOW_FIXED;
-		dev_type = E1XX;
-	} else if (e110_str != std::string::npos) {
-		tx_window = TX_WINDOW_FIXED;
-		dev_type = E1XX;
-	} else if (usrp2_str != std::string::npos) {
-		tx_window = TX_WINDOW_FIXED;
-		dev_type = USRP2;
-	} else if ((e310_str != std::string::npos) ||
-		   (e3xx_str != std::string::npos)) {
-		tx_window = TX_WINDOW_FIXED;
-		dev_type = E3XX;
-	} else if (x300_str != std::string::npos) {
-		tx_window = TX_WINDOW_FIXED;
-		dev_type = X3XX;
-	} else if (x310_str != std::string::npos) {
-		tx_window = TX_WINDOW_FIXED;
-		dev_type = X3XX;
-	} else if (umtrx_str != std::string::npos) {
-		tx_window = TX_WINDOW_FIXED;
-		dev_type = UMTRX;
-	} else if (limesdr_str != std::string::npos) {
-		tx_window = TX_WINDOW_USRP1;
-		dev_type = LIMESDR;
-	} else {
-		LOG(ALERT) << "Unknown UHD device type "
-			   << dev_str << " " << mboard_str;
-		return false;
-	}
-
-	if (tx_window == TX_WINDOW_USRP1) {
-		LOG(INFO) << "Using USRP1 type transmit window for "
-			  << dev_str << " " << mboard_str;
-	} else {
-		LOG(INFO) << "Using fixed transmit window for "
-			  << dev_str << " " << mboard_str;
-	}
 
+	// Compare UHD motherboard and device strings */
+	auto mapIter = devStringMap.begin();
+	while (mapIter != devStringMap.end()) {
+		if (devString.find(mapIter->first) != std::string::npos ||
+		    mboardString.find(mapIter->first) != std::string::npos) {
+			dev_type = std::get<0>(mapIter->second);
+			tx_window = std::get<1>(mapIter->second);
 			return true;
 		}
+		mapIter++;
+	}
+
+	LOG(ALERT) << "Unsupported device " << devString;
+	return false;
+}
 
 /*
  * Check for UHD version > 3.9.0 for E3XX support
@@ -743,6 +567,45 @@ static bool uhd_e3xx_version_chk()
 	return true;
 }
 
+void uhd_device::set_channels(bool swap)
+{
+	if (iface == MULTI_ARFCN) {
+		if (dev_type != B200 && dev_type != B210)
+			throw std::invalid_argument("Device does not support MCBTS");
+		dev_type = B2XX_MCBTS;
+		chans = 1;
+	}
+
+	if (chans > dev_param_map.at(dev_key(dev_type, tx_sps, rx_sps)).channels)
+		throw std::invalid_argument("Device does not support number of requested channels");
+
+	std::string subdev_string;
+	switch (dev_type) {
+	case B210:
+	case E3XX:
+		if (chans == 1)
+			subdev_string = swap ? "A:B" : "A:A";
+		else if (chans == 2)
+			subdev_string = swap ? "A:B A:A" : "A:A A:B";
+		break;
+	case X3XX:
+	case UMTRX:
+		if (chans == 1)
+			subdev_string = swap ? "B:0" : "A:0";
+		else if (chans == 2)
+			subdev_string = swap ? "B:0 A:0" : "A:0 B:0";
+		break;
+	default:
+		break;
+	}
+
+	if (!subdev_string.empty()) {
+		uhd::usrp::subdev_spec_t spec(subdev_string);
+		usrp_dev->set_tx_subdev_spec(spec);
+		usrp_dev->set_rx_subdev_spec(spec);
+	}
+}
+
 int uhd_device::open(const std::string &args, int ref, bool swap_channels)
 {
 	const char *refstr;
@@ -773,27 +636,10 @@ int uhd_device::open(const std::string &args, int ref, bool swap_channels)
 		return -1;
 	}
 
-	// Verify and set channels
+	try {
-	if (iface == MULTI_ARFCN) {
+		set_channels(swap_channels);
-		if ((dev_type != B200) && (dev_type != B210)) {
+        } catch (const std::exception &e) {
-			LOG(ALERT) << "Unsupported device configuration";
+		LOG(ALERT) << "Channel setting failed - " << e.what();
-			return -1;
-		}
-
-		dev_type = B2XX_MCBTS;
-		chans = 1;
-	} else if (chans == 2) {
-		if (dev_type == B210) {
-		} else if (dev_type == UMTRX) {
-			uhd::usrp::subdev_spec_t subdev_spec(swap_channels?"B:0 A:0":"A:0 B:0");
-			usrp_dev->set_tx_subdev_spec(subdev_spec);
-			usrp_dev->set_rx_subdev_spec(subdev_spec);
-		} else {
-			LOG(ALERT) << "Invalid device configuration";
-			return -1;
-		}
-	} else if (chans != 1) {
-		LOG(ALERT) << "Invalid channel combination for device";
 		return -1;
 	}
 
@@ -820,14 +666,12 @@ int uhd_device::open(const std::string &args, int ref, bool swap_channels)
 
 	usrp_dev->set_clock_source(refstr);
 
-	// Set rates
+	try {
-	double _rx_rate = select_rate(dev_type, rx_sps, iface);
+		set_rates();
-	double _tx_rate = select_rate(dev_type, tx_sps, iface);
+        } catch (const std::exception &e) {
-
+		LOG(ALERT) << "UHD rate setting failed - " << e.what();
-	if ((_tx_rate < 0.0) || (_rx_rate < 0.0))
-		return -1;
-	if (set_rates(_tx_rate, _rx_rate) < 0)
 		return -1;
+	}
 
 	// Set RF frontend bandwidth
 	if (dev_type == UMTRX) {
@@ -860,17 +704,6 @@ int uhd_device::open(const std::string &args, int ref, bool swap_channels)
 	for (size_t i = 0; i < rx_buffers.size(); i++)
 		rx_buffers[i] = new smpl_buf(buf_len, rx_rate);
 
-	// Set receive chain sample offset. Trigger the EDGE offset
-	// table by checking for 4 SPS on the receive path. No other
-	// configuration supports using 4 SPS.
-	double offset = get_dev_offset();
-	if (offset == 0.0) {
-		LOG(ERR) << "Unsupported configuration, no correction applied";
-		ts_offset = 0;
-	} else  {
-		ts_offset = (TIMESTAMP) (offset * rx_rate);
-	}
-
 	// Initialize and shadow gain values 
 	init_gains();
 
@@ -1236,7 +1069,7 @@ uhd::tune_request_t uhd_device::select_freq(double freq, size_t chan, bool tx)
 
 	/* Find center frequency between channels */
 	rf_spread = fabs(freqs[!chan] - freq);
-	if (rf_spread > B2XX_CLK_RT) {
+	if (rf_spread > dev_param_map.at(dev_key(B210, tx_sps, rx_sps)).mcr) {
 		LOG(ALERT) << rf_spread << "Hz tuning spread not supported\n";
 		return treq;
 	}
@@ -1352,7 +1185,7 @@ TIMESTAMP uhd_device::initialReadTimestamp()
 double uhd_device::fullScaleInputValue()
 {
 	if (dev_type == LIMESDR)
-		return (double) 2047 * LIMESDR_TX_AMPL;
+		return (double) SHRT_MAX * LIMESDR_TX_AMPL;
 	if (dev_type == UMTRX)
 		return (double) SHRT_MAX * UMTRX_TX_AMPL;
 	else
@@ -1361,7 +1194,6 @@ double uhd_device::fullScaleInputValue()
 
 double uhd_device::fullScaleOutputValue()
 {
-	if (dev_type == LIMESDR) return (double) 2047;
 	return (double) SHRT_MAX;
 }
 

Transceiver52M/osmo-trx-dec.cpp

@@ -1,520 +0,0 @@
-/*
- * Copyright (C) 2016-2017 Alexander Chemeris <Alexander.Chemeris@fairwaves.co>
- *
- * This library is free software; you can redistribute it and/or
- * modify it under the terms of the GNU Lesser General Public
- * License as published by the Free Software Foundation; either
- * version 2.1 of the License, or (at your option) any later version.
- *
- * This library 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
- * Lesser General Public License for more details.
- *
- * You should have received a copy of the GNU Lesser General Public
- * License along with this library; if not, write to the Free Software
- * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
- */
-
-#ifdef HAVE_CONFIG_H
-#include "config.h"
-#endif
-
-#include <limits.h>
-#include <fstream>
-#include <iomanip>
-
-#include "Logger.h"
-#include "sigProcLib.h"
-#include "signalVector.h"
-#include "Transceiver.h"
-#include "Configuration.h"
-
-extern "C" {
-#include "convolve.h"
-#include "convert.h"
-}
-
-#define DEFAULT_RX_SPS        1
-#define DEFAULT_SEARCH_WINDOW 30
-
-// Tail + data + stealing + midamble + guard (without the last 0.25)
-#define BURST_LEN_FULL        156
-// Tail + data + stealing + midamble
-#define BURST_LEN_ACTIVE      148
-// Tail + data + stealing + midamble - 2*0.5
-#define BURST_LEN_USEFUL      147
-
-// Size of a sample in bytes as stores in a file
-#define SAMPLE_SIZE_BYTES     (2 * sizeof(float))
-// Burst length in bytes as stored in a file
-#define BURST_LEN_BYTES       (BURST_LEN_FULL * SAMPLE_SIZE_BYTES)
-
-ConfigurationTable gConfig;
-
-struct trx_config {
-  std::string log_level;
-  unsigned sps;
-  unsigned tsc;
-  unsigned max_expected_delay_nb;
-  unsigned max_expected_delay_ab;
-  double full_scale;
-  bool edge;
-  CorrType type;
-  std::string filename;
-  unsigned ber_burst_avg;           ///< Average BER over this many bursts.
-                                    ///< Set to 0 to average for the whole duration.
-};
-
-class NormalBurstSoftbitMask {
-public:
-  NormalBurstSoftbitMask(SoftVector &softBits)
-  : mSoftBits(softBits)
-  {
-  }
-
-  SoftVector &bits() { return mSoftBits; }
-  SoftVector tailBitsL() { return mSoftBits.segment(0,3); }
-  SoftVector dataBitsL() { return mSoftBits.segment(3,57); }
-  SoftVector stealingBitsL() { return mSoftBits.segment(60, 1); }
-  SoftVector midambleBits() { return mSoftBits.segment(61, 26); }
-  SoftVector stealingBitsR() { return mSoftBits.segment(87, 1); }
-  SoftVector dataBitsR() { return mSoftBits.segment(88,57); }
-  SoftVector tailBitsR() { return mSoftBits.segment(145,3); }
-  SoftVector guardBits() { return mSoftBits.segment(148,8); }
-
-protected:
-  SoftVector &mSoftBits;
-};
-
-class SoftBurst {
-public:
-  SoftBurst(SoftVector *softBits, double toa=0)
-  : mSoftBits(softBits), mTOA(toa)
-  {
-    assert(mSoftBits != NULL);
-  }
-
-  ~SoftBurst()
-  {
-    delete mSoftBits;
-  }
-
-  void TOA(double TOA) { mTOA = TOA; }
-  double TOA() { return mTOA; }
-
-  NormalBurstSoftbitMask normalBurstMask() { return NormalBurstSoftbitMask(*mSoftBits); }
-
-protected:
-  SoftVector *mSoftBits;
-  double mTOA;
-};
-
-class BEREstimator {
-public:
-  BEREstimator(const PRBS& prbs)
-  : mPRBS(prbs), mTotalBits(0), mErrorBits(0), mSynchronized(false)
-  {}
-
-  unsigned synchronize(const BitVector &bits)
-  {
-    for (unsigned i=0; i<mPRBS.size(); i++) {
-      mPRBS.processBit(bits[i]);
-    }
-    mSynchronized = true;
-    return mPRBS.size();
-  }
-
-  void process(const BitVector &bits, size_t start_from = 0)
-  {
-    for (size_t i=start_from; i<bits.size(); i++) {
-      mTotalBits++;
-      if (mPRBS.generateBit() != bits.bit(i)) {
-        mErrorBits++;
-      }
-    }
-  }
-
-  void sync_and_process(const BitVector &bits)
-  {
-    unsigned skip = 0;
-    if (!mSynchronized) {
-      skip = synchronize(bits);
-    }
-
-    process(bits, skip);
-  }
-
-  void skip(size_t num)
-  {
-    for (size_t i=0; i<num; i++) {
-      mTotalBits++;
-      mErrorBits++;
-      mPRBS.generateBit();
-    }
-  }
-
-  void reset()
-  {
-    mTotalBits = 0;
-    mErrorBits = 0;
-  }
-
-  unsigned totalBits() const { return mTotalBits; }
-  unsigned errorBits() const { return mErrorBits; }
-  double BER() const { return mErrorBits/(double)mTotalBits; }
-
-  bool isSynchronized() const {return mSynchronized; }
-
-protected:
-  PRBS mPRBS;
-  unsigned mTotalBits;
-  unsigned mErrorBits;
-  bool mSynchronized;
-};
-
-double getBurstRSSI(const signalVector &burst, unsigned sps, double full_scale)
-{
-  /* Calculate average power of the burst */
-  float avg = energyDetect(burst, 20 * sps);
-  return 20.0 * log10(sqrt(avg) / full_scale);
-}
-
-void printDetectionResult(int rc)
-{
-  if (rc > 0) {
-    std::cout << "Detected correlation type: " << (CorrType)rc << std::endl;
-  } else {
-    if (rc == -SIGERR_CLIP) {
-      std::cout << "Clipping detected on received RACH or Normal Burst" << std::endl;
-    } else if (rc != SIGERR_NONE) {
-      std::cout << "Unhandled RACH or Normal Burst detection error" << std::endl;
-    } else {
-//      std::cout << "No burst detected" << std::endl;
-    }
-  }
-}
-
-SoftVector *demodulateBurst(const signalVector &burst,
-                            CorrType expected_type,
-                            unsigned sps, unsigned tsc,
-                            unsigned max_expected_delay,
-                            double &timingOffset)
-{
-  complex amp;
-  float toa;
-  int rc;
-  CorrType detected_type;
-
-  /* Detect normal or RACH bursts */
-  rc = detectAnyBurst(burst, tsc, BURST_THRESH, sps, expected_type, amp, toa,
-                      max_expected_delay);
-  printDetectionResult(rc);
-  if (rc <= 0) {
-    return NULL;
-  }
-
-  // Convert samples to symbols
-  timingOffset = toa / sps;
-  // rc > 0 means it's a detected CorrType
-  detected_type = (CorrType)rc;
-
-  return demodAnyBurst(burst, sps, amp, toa, detected_type);
-}
-
-static bool processBurst(const trx_config &config, signalVector &burst,
-                         unsigned max_expected_delay,
-                         double &RSSI,
-                         double &timingOffset,
-                         BEREstimator &berEstimator)
-{
-  RSSI = getBurstRSSI(burst, config.sps, config.full_scale);
-  SoftVector *softBits = demodulateBurst(burst, config.type, config.sps,config.tsc,
-                                         max_expected_delay, timingOffset);
-
-  /* Print burst information and content */
-  if (softBits == NULL) {
-    std::cout << "Skipped frame" << std::endl;
-    // TODO: This is different for EDGE
-    berEstimator.skip(57*2);
-    return false;
-  }
-
-  SoftBurst softBurst(softBits, timingOffset);
-  NormalBurstSoftbitMask nb = softBurst.normalBurstMask();
-
-  berEstimator.sync_and_process(nb.dataBitsL().sliced());
-  berEstimator.sync_and_process(nb.dataBitsR().sliced());
-
-  std::cout << "TOA:  " << softBurst.TOA() << " symbols" << std::endl;
-  // Exclude tail and guard bits from the energy calculation
-  std::cout << "Energy:  " << softBits->segment(3,142).getEnergy() << std::endl;
-  //std::cout << "Demodulated burst: " << *softBits << std::endl;
-  std::cout << "                  tail|--------------------------data---------------------------|f|--------midamble----------|f|--------------------------data---------------------------|tai|-guard--" << std::endl;
-  //           "                   000 010001011011110011101001100100000001010001011000100100010 0 11101111000100101110111100 0 011010111011101010011010111000101100001110101011011001011 000 1''..---"
-  std::cout << "Demodulated burst:"
-            << " " << nb.tailBitsL()
-            << " " << nb.dataBitsL()
-            << " " << nb.stealingBitsL()
-            << " " << nb.midambleBits()
-            << " " << nb.stealingBitsR()
-            << " " << nb.dataBitsR()
-            << " " << nb.tailBitsR()
-            << " " << nb.guardBits()
-            << std::endl;
-
-  return true;
-}
-
-// Setup configuration values
-static void print_config(struct trx_config *config)
-{
-  std::ostringstream ost("");
-  ost << "Config Settings" << std::endl;
-  ost << "   Source file name............. " << config->filename << std::endl;
-  ost << "   Log Level.................... " << config->log_level << std::endl;
-  ost << "   Rx Samples-per-Symbol........ " << config->sps << std::endl;
-  ost << "   EDGE support................. " << (config->edge ? "Enabled" : "Disabled") << std::endl;
-  ost << "   Burst type................... " << config->type << std::endl;
-  ost << "   Burst TSC.................... " << config->tsc << std::endl;
-  ost << "   Normal Burst search window... " << config->max_expected_delay_nb << std::endl;
-  ost << "   Access Burst search window... " << config->max_expected_delay_ab << std::endl;
-  ost << "   Signal full scale............ " << config->full_scale << std::endl;
-  ost << "   BER average window (bursts).. " << config->ber_burst_avg << std::endl;
-  std::cout << ost << std::endl;
-}
-
-static void print_help()
-{
-  fprintf(stdout, "Options:\n"
-          "  -h          This text\n"
-          "  -l LEVEL    Logging level (%s)\n"
-          "  -e          Enable EDGE receiver\n"
-          "  -s SPS      Samples-per-symbol (1 or 4, default: %d)\n"
-          "  -t TSC      Burst training sequence (0 to 7, default: 0)\n"
-          "  -f FILE     File to read\n"
-          "  -w SYMBOLS  Normal Burst search window (0 to 156, default: %d)\n"
-          "  -W SYMBOLS  Access Burst search window (0 to 156, default: %d)\n"
-          "  -b BURSTS   BER average window. Set to 0 to average over the whole file (default: 1)\n",
-          "EMERG, ALERT, CRT, ERR, WARNING, NOTICE, INFO, DEBUG",
-          DEFAULT_RX_SPS,
-          DEFAULT_SEARCH_WINDOW, DEFAULT_SEARCH_WINDOW);
-}
-
-static bool handle_options(int argc, char **argv, struct trx_config *config)
-{
-  int option;
-
-  config->log_level = "NOTICE";
-  config->sps = DEFAULT_RX_SPS;
-  config->tsc = 0;
-  config->max_expected_delay_nb = DEFAULT_SEARCH_WINDOW;
-  config->max_expected_delay_ab = DEFAULT_SEARCH_WINDOW;
-  config->full_scale = SHRT_MAX;
-  config->edge = false;
-  config->type = TSC;
-  config->ber_burst_avg = 1;
-
-  while ((option = getopt(argc, argv, "ls:et:f:w:W:b:h")) != -1) {
-    switch (option) {
-    case 'l':
-      config->log_level = optarg;
-      break;
-    case 's':
-      config->sps = atoi(optarg);
-      break;
-    case 'e':
-      config->edge = true;
-      break;
-    case 't':
-      config->tsc = atoi(optarg);
-      break;
-    case 'f':
-      config->filename = optarg;
-      break;
-    case 'w':
-      config->max_expected_delay_nb = atoi(optarg);
-      break;
-    case 'W':
-      config->max_expected_delay_ab = atoi(optarg);
-      break;
-    case 'b':
-      config->ber_burst_avg = atoi(optarg);
-      break;
-    case 'h':
-    default:
-      print_help();
-      exit(0);
-    }
-  }
-
-  if ((config->sps != 1) && (config->sps != 4)) {
-    printf("ERROR: Unsupported samples-per-symbol %i\n\n", config->sps);
-    return false;
-  }
-
-  if (config->edge && (config->sps != 4)) {
-    printf("ERROR: EDGE only supported at 4 samples per symbol\n\n");
-    return false;
-  }
-
-  if (config->tsc > 7) {
-    printf("ERROR: Invalid training sequence %i\n\n", config->tsc);
-    return false;
-  }
-
-  if (config->filename.length() == 0) {
-    printf("ERROR: No input file specified\n\n");
-    return false;
-  }
-
-  if (config->max_expected_delay_nb > 156 || config->max_expected_delay_nb < 0 ||
-      config->max_expected_delay_ab > 156 || config->max_expected_delay_ab < 0) {
-    printf("ERROR: Invalid search window size, must be withit [1..156] range\n\n");
-    return false;
-  }
-
-  return true;
-}
-
-int main(int argc, char *argv[])
-{
-  struct trx_config config;
-
-#ifdef HAVE_SSE3
-  printf("Info: SSE3 support compiled in");
-  if (__builtin_cpu_supports("sse3"))
-    printf(" and supported by CPU\n");
-  else
-    printf(", but not supported by CPU\n");
-#endif
-
-#ifdef HAVE_SSE4_1
-  printf("Info: SSE4.1 support compiled in");
-  if (__builtin_cpu_supports("sse4.1"))
-    printf(" and supported by CPU\n");
-  else
-    printf(", but not supported by CPU\n");
-#endif
-
-  convolve_init();
-  convert_init();
-
-  // Process command line options and print config to screen
-  if (!handle_options(argc, argv, &config)) {
-    print_help();
-    exit(0);
-  }
-  print_config(&config);
-
-  gLogInit("transceiver", config.log_level.c_str(), LOG_LOCAL7);
-
-  if (!sigProcLibSetup()) {
-    LOG(ALERT) << "Failed to initialize signal processing library";
-    return -1;
-  }
-
-  double RSSI;
-  double timingOffset, timingOffsetPrev = 0.0;
-  signalVector burst(2*BURST_LEN_FULL);
-  GSM::Time gsmTime;
-  bool syncedTo157bits = false; // We should syncronize to 156-157 frame structure only once
-  bool burst156_157 = false;    // Set to true to enable 156-156-156-157 frame
-  int bitsReadExtra = 0; // set to 1 every 4 bursts and when TOA>1.0
-  int bitsToSkip = 0;    // set to 1 when TOA<0.0
-  unsigned berBurstsAveraged = 0;
-  PRBS9 prbs;
-  BEREstimator berEstimator(prbs);
-
-  // Configure output stream
-  std::cout << std::fixed;
-  std::cout << std::setprecision(2);
-
-  std::ifstream file (config.filename.c_str(), std::ifstream::binary);
-
-  // Read the first burst, but do not process it, because we need at least two bursts
-  // worth of data for reliable initial detection.
-  file.read((char*)burst.begin(), config.sps * BURST_LEN_BYTES);
-  {signalVector t = burst.segment(0, BURST_LEN_FULL); scaleVector(t, complex(SHRT_MAX)); }
-
-#if 0
-    /* Distort signal */
-    {
-      signalVector burst_read = burst.segment(85,156);
-      std::ifstream file (config.filename.c_str(), std::ifstream::binary);
-      file.read((char*)burst_read.begin(), burst_read.size() * 2 * sizeof(float));
-      file.close();
-    }
-#endif
-#if 1
-  // Read more data and try burst detection until successful
-  while(file.read((char*)(burst.begin()+config.sps*BURST_LEN_FULL), config.sps*BURST_LEN_BYTES))
-  {
-    {signalVector t = burst.segment(BURST_LEN_FULL, BURST_LEN_FULL); scaleVector(t, complex(SHRT_MAX)); }
-    bool found = processBurst(config, burst, BURST_LEN_FULL, RSSI, timingOffset, berEstimator);
-    std::cout << "RSSI: " << RSSI << " dBFS" << std::endl;
-    if (found) {
-      gsmTime.incTN();
-      berBurstsAveraged++;
-      break;
-    }
-    burst.segmentMove(config.sps*BURST_LEN_FULL, 0, config.sps*BURST_LEN_FULL);
-  }
-
-  // Align stream to burst
-  int offsetInt = (int)timingOffset;
-  burst.segmentMove(config.sps*(BURST_LEN_FULL+offsetInt), 0, config.sps*(BURST_LEN_FULL-offsetInt));
-  {signalVector t = burst.segment(0, BURST_LEN_FULL-offsetInt); scaleVector(t, complex(1.0/SHRT_MAX)); }
-  file.read((char*)(burst.begin()+config.sps*(BURST_LEN_FULL-offsetInt)), config.sps*offsetInt*SAMPLE_SIZE_BYTES);
-#endif
-
-  // Resize burst vector to hold only one burst, because demodulation code
-  // always decode the full vector size.
-  burst.shrink(BURST_LEN_FULL+1);
-
-  // Process the rest of the stream
-  do {
-    {signalVector t = burst.segment(0, BURST_LEN_FULL); scaleVector(t, complex(SHRT_MAX)); }
-    processBurst(config, burst, (config.type==RACH)?config.max_expected_delay_ab:config.max_expected_delay_ab,
-                 RSSI, timingOffset, berEstimator);
-    if (burst156_157 && !syncedTo157bits && timingOffset - timingOffsetPrev > .75) {
-      std::cout << "TOA adjust: Found a 157-bit burst, reset TN to mark it" << std::endl;
-      gsmTime.TN(2);
-      timingOffset -= 1.0;
-      // Make sure we do this adjustment only once.
-      syncedTo157bits = true;
-    } else {
-      gsmTime.incTN();
-    }
-    bitsToSkip = 0;
-    bitsReadExtra = 0;
-    if (timingOffset < 0.0) {
-      std::cout << "TOA adjust: skip a bit" << std::endl;
-      burst[0] = 0;
-      bitsToSkip = 1;
-      bitsReadExtra--;
-    }
-    bitsReadExtra += (gsmTime.TN()%4 == 0);
-    if (timingOffset > 1.1) {
-      std::cout << "TOA adjust: add extra bit" << std::endl;
-      bitsReadExtra++;
-    }
-    std::cout << "Clock: " << gsmTime;
-    std::cout << " RSSI: " << RSSI << " dBFS";
-    std::cout << " Error bits: " << berEstimator.errorBits() << " Total bits: " << berEstimator.totalBits()
-              << " BER: " << 100.0*berEstimator.errorBits() / berEstimator.totalBits() << "%" << std::endl;
-    berBurstsAveraged++;
-    // Never reset if config.ber_burst_avg is 0
-    if (config.ber_burst_avg > 0 && berBurstsAveraged >= config.ber_burst_avg) {
-        berBurstsAveraged = 0;
-        berEstimator.reset();
-    }
-    std::cout << "bitsReadExtra: " << bitsReadExtra << " bitsToSkip: " << bitsToSkip << std::endl;
-    timingOffsetPrev = timingOffset;
-  } while(file.read((char*)(burst.begin()+bitsToSkip), config.sps*(BURST_LEN_BYTES+SAMPLE_SIZE_BYTES*bitsReadExtra)));
-
-  std::cout << "End of file reached" << std::endl;
-  file.close();
-
-  return 0;
-}

Transceiver52M/osmo-trx-gen.cpp

@@ -1,334 +0,0 @@
-/*
- * Copyright (C) 2017 Alexander Chemeris <Alexander.Chemeris@fairwaves.co>
- *
- * This library is free software; you can redistribute it and/or
- * modify it under the terms of the GNU Lesser General Public
- * License as published by the Free Software Foundation; either
- * version 2.1 of the License, or (at your option) any later version.
- *
- * This library 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
- * Lesser General Public License for more details.
- *
- * You should have received a copy of the GNU Lesser General Public
- * License along with this library; if not, write to the Free Software
- * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
- */
-
-#ifdef HAVE_CONFIG_H
-#include "config.h"
-#endif
-
-#include <limits.h>
-#include <fstream>
-#include <iomanip>
-#include <endian.h> // for byte order manipulation
-
-#include "Logger.h"
-#include "sigProcLib.h"
-#include "GSMCommon.h"
-#include "BitVector.h"
-#include "Configuration.h"
-
-extern "C" {
-#include "convolve.h"
-#include "convert.h"
-}
-
-#define DEFAULT_SPS           4
-#define DEFAULT_SEARCH_WINDOW 30
-
-// Tail + data + stealing + midamble + guard (without the last 0.25)
-#define BURST_LEN_FULL        156
-// Tail + data + stealing + midamble
-#define BURST_LEN_ACTIVE      148
-// Tail + data + stealing + midamble - 2*0.5
-#define BURST_LEN_USEFUL      147
-
-// Size of a sample in bytes as stores in a file
-#define SAMPLE_SIZE_BYTES     (2 * sizeof(float))
-// Burst length in bytes as stored in a file
-#define BURST_LEN_BYTES       (BURST_LEN_FULL * SAMPLE_SIZE_BYTES)
-
-ConfigurationTable gConfig;
-
-enum FileType {
-  FLOAT_NORM_LE,  ///< Float -1..+1 Little Endian
-  FLOAT16_LE,     ///< Float -32767..+32767 Little Endian
-  SIGNED16_LE,    ///< Integer -32767..+32767 Little Endian
-  SIGNED16_BE,    ///< Integer -32767..+32767 Big Endian (Keysight waveform format)
-};
-
-struct trx_config {
-  std::string log_level;
-  unsigned sps;
-  unsigned tsc;
-  double full_scale;
-  bool edge;
-  CorrType type;
-  std::string filename;
-  FileType file_type;
-};
-
-std::ostream& operator<<(std::ostream& os, FileType ftype)
-{
-  switch(ftype)
-  {
-  case FLOAT_NORM_LE:
-    os << "float";
-    break;
-  case FLOAT16_LE:
-    os << "float16";
-    break;
-  case SIGNED16_LE:
-    os << "signed16";
-    break;
-  case SIGNED16_BE:
-    os << "signed16be";
-    break;
-  default:
-    assert(!"unknown file type");
-  }
-	return os;
-}
-
-
-void writeBurstFloatNorm(std::ofstream& os, const signalVector& v)
-{
-  os.write((char*)v.begin(), v.size() * 2 * sizeof(float));
-}
-
-void writeBurstFloat16LE(std::ofstream& os, const signalVector& v)
-{
-  const complex *c = v.begin();
-  for (size_t i=0; i<v.size(); i++, c++) {
-    float iq[2];
-    iq[0] = c->real()*SHRT_MAX;
-    iq[1] = c->imag()*SHRT_MAX;
-    os.write((char*)&iq, 2*sizeof(float));
-  }
-}
-
-void writeBurstSigned16LE(std::ofstream& os, const signalVector& v)
-{
-  const complex *c = v.begin();
-  for (size_t i=0; i<v.size(); i++, c++) {
-    int16_t iq[2];
-    iq[0] = c->real()*SHRT_MAX;
-    iq[1] = c->imag()*SHRT_MAX;
-    iq[0] = htole16(iq[0]);
-    iq[1] = htole16(iq[1]);
-    os.write((char*)&iq, 2*sizeof(int16_t));
-  }
-}
-
-void writeBurstSigned16BE(std::ofstream& os, const signalVector& v)
-{
-  const complex *c = v.begin();
-  for (size_t i=0; i<v.size(); i++, c++) {
-    int16_t iq[2];
-    iq[0] = c->real()*SHRT_MAX;
-    iq[1] = c->imag()*SHRT_MAX;
-    iq[0] = htobe16(iq[0]);
-    iq[1] = htobe16(iq[1]);
-    os.write((char*)&iq, 2*sizeof(int16_t));
-  }
-}
-
-void writeBurst(std::ofstream& os, const signalVector& v, FileType ftype)
-{
-  switch(ftype)
-  {
-  case FLOAT_NORM_LE:
-    writeBurstFloatNorm(os, v);
-    break;
-  case FLOAT16_LE:
-    writeBurstFloat16LE(os, v);
-    break;
-  case SIGNED16_LE:
-    writeBurstSigned16LE(os, v);
-    break;
-  case SIGNED16_BE:
-    writeBurstSigned16BE(os, v);
-    break;
-  default:
-    assert(!"unknown file type");
-  }
-}
-
-// Setup configuration values
-static void print_config(struct trx_config *config)
-{
-  std::ostringstream ost("");
-  ost << "Config Settings" << std::endl;
-  ost << "   Destination file name........ " << config->filename << std::endl;
-  ost << "   Destination file type........ " << config->file_type << std::endl;
-  ost << "   Log Level.................... " << config->log_level << std::endl;
-  ost << "   Tx Samples-per-Symbol........ " << config->sps << std::endl;
-  ost << "   EDGE support................. " << (config->edge ? "Enabled" : "Disabled") << std::endl;
-  ost << "   Burst type................... " << config->type << std::endl;
-  ost << "   Burst TSC.................... " << config->tsc << std::endl;
-  ost << "   Signal full scale............ " << config->full_scale << std::endl;
-  std::cout << ost << std::endl;
-}
-
-static void print_help()
-{
-  fprintf(stdout,
-          "This utility generates waveform files aka IQ binary files in a number of formats"
-          "to use them as input to osmo-trx-dec or load them into signal generators.\n"
-          "\n"
-          "Options:\n"
-          "  -h          This text\n"
-          "  -l LEVEL    Logging level (%s)\n"
-          "  -e          Enable EDGE receiver\n"
-          "  -s SPS      Samples-per-symbol (1 or 4, default: %d)\n"
-          "  -t TSC      Burst training sequence (0 to 7, default: 0)\n"
-          "  -f FILE     File to write generated bursts to\n"
-          "  -F FILETYPE Format of the file - float, float16, signed16, signed16be (default: f16)\n"
-          "              Note: Keysight waveform format is signed16be. osmo-trx-dec accepts float16.\n",
-          "EMERG, ALERT, CRT, ERR, WARNING, NOTICE, INFO, DEBUG",
-          DEFAULT_SPS);
-}
-
-FileType option_to_file_type(const std::string &optarg)
-{
-  if (optarg == "float") {
-    return FLOAT_NORM_LE;
-  } else if (optarg == "float16") {
-    return FLOAT16_LE;
-  } else if (optarg == "signed16") {
-    return SIGNED16_LE;
-  } else if (optarg == "signed16be") {
-    return SIGNED16_BE;
-  } else {
-    return (FileType)-1;
-  }
-}
-
-static bool handle_options(int argc, char **argv, struct trx_config *config)
-{
-  int option;
-
-  config->log_level = "NOTICE";
-  config->sps = DEFAULT_SPS;
-  config->tsc = 0;
-  config->full_scale = SHRT_MAX;
-  config->edge = false;
-  config->type = TSC;
-  config->file_type = FLOAT16_LE;
-
-  while ((option = getopt(argc, argv, "ls:et:f:F:h")) != -1) {
-    switch (option) {
-    case 'l':
-      config->log_level = optarg;
-      break;
-    case 's':
-      config->sps = atoi(optarg);
-      break;
-    case 'e':
-      config->edge = true;
-      break;
-    case 't':
-      config->tsc = atoi(optarg);
-      break;
-    case 'f':
-      config->filename = optarg;
-      break;
-    case 'F':
-      config->file_type = option_to_file_type(optarg);
-      break;
-    case 'h':
-    default:
-      print_help();
-      exit(0);
-    }
-  }
-
-  if ((config->sps != 1) && (config->sps != 4)) {
-    printf("ERROR: Unsupported samples-per-symbol %i\n\n", config->sps);
-    return false;
-  }
-
-  if (config->edge && (config->sps != 4)) {
-    printf("ERROR: EDGE only supported at 4 samples per symbol\n\n");
-    return false;
-  }
-
-  if (config->tsc > 7) {
-    printf("ERROR: Invalid training sequence %i\n\n", config->tsc);
-    return false;
-  }
-
-  if (config->filename.length() == 0) {
-    printf("ERROR: No output file name specified\n\n");
-    return false;
-  }
-
-  if (config->file_type < 0) {
-    printf("ERROR: Wrong output file format\n\n");
-  }
-
-  return true;
-}
-
-int main(int argc, char *argv[])
-{
-  struct trx_config config;
-
-#ifdef HAVE_SSE3
-  printf("Info: SSE3 support compiled in");
-  if (__builtin_cpu_supports("sse3"))
-    printf(" and supported by CPU\n");
-  else
-    printf(", but not supported by CPU\n");
-#endif
-
-#ifdef HAVE_SSE4_1
-  printf("Info: SSE4.1 support compiled in");
-  if (__builtin_cpu_supports("sse4.1"))
-    printf(" and supported by CPU\n");
-  else
-    printf(", but not supported by CPU\n");
-#endif
-
-  convolve_init();
-  convert_init();
-
-  // Process command line options and print config to screen
-  if (!handle_options(argc, argv, &config)) {
-    print_help();
-    exit(0);
-  }
-  print_config(&config);
-
-  gLogInit("transceiver", config.log_level.c_str(), LOG_LOCAL7);
-
-  if (!sigProcLibSetup()) {
-    LOG(ALERT) << "Failed to initialize signal processing library";
-    return -1;
-  }
-
-  signalVector burst(2*BURST_LEN_FULL);
-  GSM::Time gsmTime;
-  PRBS9 prbs;
-
-  // Configure output stream
-  std::cout << std::fixed;
-  std::cout << std::setprecision(2);
-
-  std::ofstream file (config.filename.c_str(), std::ifstream::binary);
-
-  for (int i=0; i<511; i++) {
-    signalVector *signal = genRandNormalBurst(config.tsc, config.sps, gsmTime.TN(), prbs);
-    writeBurst(file, *signal, config.file_type);
-    gsmTime.incTN();
-  }
-
-  file.close();
-
-  std::cout << "Done!" << std::endl;
-
-  return 0;
-}

Transceiver52M/osmo-trx.cpp

@@ -27,6 +27,7 @@
 #include <signal.h>
 #include <stdlib.h>
 #include <unistd.h>
+#include <sched.h>
 
 #include <GSMCommon.h>
 #include <Logger.h>
@@ -61,7 +62,8 @@ extern "C" {
 
 struct trx_config {
 	std::string log_level;
-	std::string addr;
+	std::string local_addr;
+	std::string remote_addr;
 	std::string dev_args;
 	unsigned port;
 	unsigned tx_sps;
@@ -77,6 +79,7 @@ struct trx_config {
 	double rssi_offset;
 	bool swap_channels;
 	bool edge;
+	int sched_rr;
 };
 
 ConfigurationTable gConfig;
@@ -132,7 +135,8 @@ bool trx_setup_config(struct trx_config *config)
 	ost << "   Log Level............... " << config->log_level << std::endl;
 	ost << "   Device args............. " << config->dev_args << std::endl;
 	ost << "   TRX Base Port........... " << config->port << std::endl;
-	ost << "   TRX Address............. " << config->addr << std::endl;
+	ost << "   TRX Address............. " << config->local_addr << std::endl;
+	ost << "   GSM Core Address........." << config->remote_addr << std::endl;
 	ost << "   Channels................ " << config->chans << std::endl;
 	ost << "   Tx Samples-per-Symbol... " << config->tx_sps << std::endl;
 	ost << "   Rx Samples-per-Symbol... " << config->rx_sps << std::endl;
@@ -198,9 +202,10 @@ Transceiver *makeTransceiver(struct trx_config *config, RadioInterface *radio)
 	Transceiver *trx;
 	VectorFIFO *fifo;
 
-	trx = new Transceiver(config->port, config->addr.c_str(),
+	trx = new Transceiver(config->port, config->local_addr.c_str(),
-			      config->tx_sps, config->rx_sps, config->chans,
+			      config->remote_addr.c_str(), config->tx_sps,
-			      GSM::Time(3,0), radio, config->rssi_offset);
+			      config->rx_sps, config->chans, GSM::Time(3,0),
+			      radio, config->rssi_offset);
 	if (!trx->init(config->filler, config->rtsc,
 		       config->rach_delay, config->edge)) {
 		LOG(ALERT) << "Failed to initialize transceiver";
@@ -246,6 +251,7 @@ static void print_help()
 		"  -a    UHD device args\n"
 		"  -l    Logging level (%s)\n"
 		"  -i    IP address of GSM core\n"
+		"  -j    IP address of osmo-trx\n"
 		"  -p    Base port number\n"
 		"  -e    Enable EDGE receiver\n"
 		"  -m    Enable multi-ARFCN transceiver (default=disabled)\n"
@@ -256,11 +262,11 @@ static void print_help()
 		"  -c    Number of ARFCN channels (default=1)\n"
 		"  -f    Enable C0 filler table\n"
 		"  -o    Set baseband frequency offset (default=auto)\n"
-		"  -r    Random GMSK Normal Burst test mode with given TSC\n"
+		"  -r    Random Normal Burst test mode with TSC\n"
-		"  -E    Random 8-PSK Normal Burst test mode with given TSC\n"
 		"  -A    Random Access Burst test mode with delay\n"
 		"  -R    RSSI to dBm offset in dB (default=0)\n"
-		"  -S    Swap channels (UmTRX only)\n",
+		"  -S    Swap channels (UmTRX only)\n"
+		"  -t    SCHED_RR real-time priority (1..32)\n",
 		"EMERG, ALERT, CRT, ERR, WARNING, NOTICE, INFO, DEBUG");
 }
 
@@ -269,7 +275,8 @@ static void handle_options(int argc, char **argv, struct trx_config *config)
 	int option;
 
 	config->log_level = "NOTICE";
-	config->addr = DEFAULT_TRX_IP;
+	config->local_addr = DEFAULT_TRX_IP;
+	config->remote_addr = DEFAULT_TRX_IP;
 	config->port = DEFAULT_TRX_PORT;
 	config->tx_sps = DEFAULT_TX_SPS;
 	config->rx_sps = DEFAULT_RX_SPS;
@@ -284,8 +291,9 @@ static void handle_options(int argc, char **argv, struct trx_config *config)
 	config->rssi_offset = 0.0;
 	config->swap_channels = false;
 	config->edge = false;
+	config->sched_rr = -1;
 
-	while ((option = getopt(argc, argv, "ha:l:i:p:c:dmxgfo:s:b:r:E:A:R:Se")) != -1) {
+	while ((option = getopt(argc, argv, "ha:l:i:j:p:c:dmxgfo:s:b:r:A:R:Set:")) != -1) {
 		switch (option) {
 		case 'h':
 			print_help();
@@ -298,7 +306,10 @@ static void handle_options(int argc, char **argv, struct trx_config *config)
 			config->log_level = optarg;
 			break;
 		case 'i':
-			config->addr = optarg;
+			config->remote_addr = optarg;
+			break;
+		case 'j':
+			config->local_addr = optarg;
 			break;
 		case 'p':
 			config->port = atoi(optarg);
@@ -331,10 +342,6 @@ static void handle_options(int argc, char **argv, struct trx_config *config)
 			config->rtsc = atoi(optarg);
 			config->filler = Transceiver::FILLER_NORM_RAND;
 			break;
-		case 'E':
-			config->rtsc = atoi(optarg);
-			config->filler = Transceiver::FILLER_EDGE_RAND;
-			break;
 		case 'A':
 			config->rach_delay = atoi(optarg);
 			config->filler = Transceiver::FILLER_ACCESS_RAND;
@@ -348,6 +355,9 @@ static void handle_options(int argc, char **argv, struct trx_config *config)
 		case 'e':
 			config->edge = true;
 			break;
+		case 't':
+			config->sched_rr = atoi(optarg);
+			break;
 		default:
 			print_help();
 			exit(0);
@@ -365,10 +375,8 @@ static void handle_options(int argc, char **argv, struct trx_config *config)
 		goto bad_config;
 	}
 
-	if (!config->edge && (config->filler == Transceiver::FILLER_EDGE_RAND)) {
+	if (config->edge && (config->filler == Transceiver::FILLER_NORM_RAND))
-		printf("Can't enable EDGE filler when EDGE mode is disabled\n\n");
+		config->filler = Transceiver::FILLER_EDGE_RAND;
-		goto bad_config;
-	}
 
 	if ((config->tx_sps != 1) && (config->tx_sps != 4) &&
 	    (config->rx_sps != 1) && (config->rx_sps != 4)) {
@@ -393,6 +401,21 @@ bad_config:
 	exit(0);
 }
 
+static int set_sched_rr(int prio)
+{
+	struct sched_param param;
+	int rc;
+	memset(&param, 0, sizeof(param));
+	param.sched_priority = prio;
+	printf("Setting SCHED_RR priority(%d)\n", param.sched_priority);
+	rc = sched_setscheduler(getpid(), SCHED_RR, &param);
+	if (rc != 0) {
+		std::cerr << "Config: Setting SCHED_RR failed" << std::endl;
+		return -1;
+	}
+	return 0;
+}
+
 int main(int argc, char *argv[])
 {
 	int type, chans, ref;
@@ -404,18 +427,26 @@ int main(int argc, char *argv[])
 
 #ifdef HAVE_SSE3
 	printf("Info: SSE3 support compiled in");
+#ifdef HAVE___BUILTIN_CPU_SUPPORTS
 	if (__builtin_cpu_supports("sse3"))
 		printf(" and supported by CPU\n");
 	else
 		printf(", but not supported by CPU\n");
+#else
+	printf(", but runtime SIMD detection disabled\n");
+#endif
 #endif
 
 #ifdef HAVE_SSE4_1
 	printf("Info: SSE4.1 support compiled in");
+#ifdef HAVE___BUILTIN_CPU_SUPPORTS
 	if (__builtin_cpu_supports("sse4.1"))
 		printf(" and supported by CPU\n");
 	else
 		printf(", but not supported by CPU\n");
+#else
+	printf(", but runtime SIMD detection disabled\n");
+#endif
 #endif
 
 	convolve_init();
@@ -423,6 +454,11 @@ int main(int argc, char *argv[])
 
 	handle_options(argc, argv, &config);
 
+	if (config.sched_rr != -1) {
+		if (set_sched_rr(config.sched_rr) < 0)
+			return EXIT_FAILURE;
+	}
+
 	setup_signal_handlers();
 
 	/* Check database sanity */

Transceiver52M/radioInterface.cpp

@@ -22,7 +22,6 @@
 #include "radioInterface.h"
 #include "Resampler.h"
 #include <Logger.h>
-#include <PRBS.h>
 
 extern "C" {
 #include "convert.h"

Transceiver52M/sigProcLib.cpp

@@ -46,14 +46,10 @@ using namespace GSM;
 #define CLIP_THRESH		30000.0f
 
 /** Lookup tables for trigonometric approximation */
-float cosTable[TABLESIZE+1]; // add 1 element for wrap around
+static float sincTable[TABLESIZE+1]; // add 1 element for wrap around
-float sinTable[TABLESIZE+1];
-float sincTable[TABLESIZE+1];
 
 /** Constants */
 static const float M_PI_F = (float)M_PI;
-static const float M_2PI_F = (float)(2.0*M_PI);
-static const float M_1_2PI_F = 1/M_2PI_F;
 
 /* Precomputed rotation vectors */
 static signalVector *GMSKRotation4 = NULL;
@@ -64,7 +60,7 @@ static signalVector *GMSKReverseRotation1 = NULL;
 /* Precomputed fractional delay filters */
 static signalVector *delayFilters[DELAYFILTS];
 
-static Complex<float> psk8_table[8] = {
+static const Complex<float> psk8_table[8] = {
    Complex<float>(-0.70710678,  0.70710678),
    Complex<float>( 0.0, -1.0),
    Complex<float>( 0.0,  1.0),
@@ -172,67 +168,7 @@ void sigProcLibDestroy()
   GSMPulse4 = NULL;
 }
 
-// dB relative to 1.0.
+static float vectorNorm2(const signalVector &x)
-// if > 1.0, then return 0 dB
-float dB(float x) {
-  
-  float arg = 1.0F;
-  float dB = 0.0F;
-  
-  if (x >= 1.0F) return 0.0F;
-  if (x <= 0.0F) return -200.0F;
-
-  float prevArg = arg;
-  float prevdB = dB;
-  float stepSize = 16.0F;
-  float dBstepSize = 12.0F;
-  while (stepSize > 1.0F) {
-    do {
-      prevArg = arg;
-      prevdB = dB;
-      arg /= stepSize;
-      dB -= dBstepSize;
-    } while (arg > x);
-    arg = prevArg;
-    dB = prevdB;
-    stepSize *= 0.5F;
-    dBstepSize -= 3.0F;
-  }
- return ((arg-x)*(dB-3.0F) + (x-arg*0.5F)*dB)/(arg - arg*0.5F);
-
-}
-
-// 10^(-dB/10), inverse of dB func.
-float dBinv(float x) {
-  
-  float arg = 1.0F;
-  float dB = 0.0F;
-  
-  if (x >= 0.0F) return 1.0F;
-  if (x <= -200.0F) return 0.0F;
-
-  float prevArg = arg;
-  float prevdB = dB;
-  float stepSize = 16.0F;
-  float dBstepSize = 12.0F;
-  while (stepSize > 1.0F) {
-    do {
-      prevArg = arg;
-      prevdB = dB;
-      arg /= stepSize;
-      dB -= dBstepSize;
-    } while (dB > x);
-    arg = prevArg;
-    dB = prevdB;
-    stepSize *= 0.5F;
-    dBstepSize -= 3.0F;
-  }
-
-  return ((dB-x)*(arg*0.5F)+(x-(dB-3.0F))*(arg))/3.0F;
-
-}
-
-float vectorNorm2(const signalVector &x) 
 {
   signalVector::const_iterator xPtr = x.begin();
   float Energy = 0.0;
@@ -242,64 +178,6 @@ float vectorNorm2(const signalVector &x)
   return Energy;
 }
 
-
-float vectorPower(const signalVector &x) 
-{
-  return vectorNorm2(x)/x.size();
-}
-
-/** compute cosine via lookup table */
-float cosLookup(const float x)
-{
-  float arg = x*M_1_2PI_F;
-  while (arg > 1.0F) arg -= 1.0F;
-  while (arg < 0.0F) arg += 1.0F;
-
-  const float argT = arg*((float)TABLESIZE);
-  const int argI = (int)argT;
-  const float delta = argT-argI;
-  const float iDelta = 1.0F-delta;
-  return iDelta*cosTable[argI] + delta*cosTable[argI+1];
-}
-
-/** compute sine via lookup table */
-float sinLookup(const float x) 
-{
-  float arg = x*M_1_2PI_F;
-  while (arg > 1.0F) arg -= 1.0F;
-  while (arg < 0.0F) arg += 1.0F;
-
-  const float argT = arg*((float)TABLESIZE);
-  const int argI = (int)argT;
-  const float delta = argT-argI;
-  const float iDelta = 1.0F-delta;
-  return iDelta*sinTable[argI] + delta*sinTable[argI+1];
-}
-
-
-/** compute e^(-jx) via lookup table. */
-static complex expjLookup(float x)
-{
-  float arg = x*M_1_2PI_F;
-  while (arg > 1.0F) arg -= 1.0F;
-  while (arg < 0.0F) arg += 1.0F;
-
-  const float argT = arg*((float)TABLESIZE);
-  const int argI = (int)argT;
-  const float delta = argT-argI;
-  const float iDelta = 1.0F-delta;
-   return complex(iDelta*cosTable[argI] + delta*cosTable[argI+1],
-		   iDelta*sinTable[argI] + delta*sinTable[argI+1]);
-}
-
-/** Library setup functions */
-static void initTrigTables() {
-  for (int i = 0; i < TABLESIZE+1; i++) {
-    cosTable[i] = cos(2.0*M_PI*i/TABLESIZE);
-    sinTable[i] = sin(2.0*M_PI*i/TABLESIZE);
-  }
-}
-
 /*
  * Initialize 4 sps and 1 sps rotation tables
  */
@@ -311,11 +189,11 @@ static void initGMSKRotationTables()
   GMSKReverseRotation4 = new signalVector(len4);
   signalVector::iterator rotPtr = GMSKRotation4->begin();
   signalVector::iterator revPtr = GMSKReverseRotation4->begin();
-  float phase = 0.0;
+  auto phase = 0.0;
   while (rotPtr != GMSKRotation4->end()) {
-    *rotPtr++ = expjLookup(phase);
+    *rotPtr++ = complex(cos(phase), sin(phase));
-    *revPtr++ = expjLookup(-phase);
+    *revPtr++ = complex(cos(-phase), sin(-phase));
-    phase += M_PI_F / 2.0F / 4.0;
+    phase += M_PI / 2.0 / 4.0;
   }
 
   GMSKRotation1 = new signalVector(len1);
@@ -324,9 +202,9 @@ static void initGMSKRotationTables()
   revPtr = GMSKReverseRotation1->begin();
   phase = 0.0;
   while (rotPtr != GMSKRotation1->end()) {
-    *rotPtr++ = expjLookup(phase);
+    *rotPtr++ = complex(cos(phase), sin(phase));
-    *revPtr++ = expjLookup(-phase);
+    *revPtr++ = complex(cos(-phase), sin(-phase));
-    phase += M_PI_F / 2.0F;
+    phase += M_PI / 2.0;
   }
 }
 
@@ -401,11 +279,18 @@ static bool GMSKReverseRotate(signalVector &x, int sps)
   return true;
 }
 
-signalVector *convolve(const signalVector *x,
+/** Convolution type indicator */
-                        const signalVector *h,
+enum ConvType {
-                        signalVector *y,
+  START_ONLY,
-                        ConvType spanType, size_t start,
+  NO_DELAY,
-                        size_t len, size_t step, int offset)
+  CUSTOM,
+  UNDEFINED,
+};
+
+static signalVector *convolve(const signalVector *x, const signalVector *h,
+                              signalVector *y, ConvType spanType,
+                              size_t start = 0, size_t len = 0,
+                              size_t step = 1, int offset = 0)
 {
   int rc;
   size_t head = 0, tail = 0;
@@ -654,29 +539,6 @@ static PulseSequence *generateGSMPulse(int sps)
   return pulse;
 }
 
-signalVector* reverseConjugate(signalVector *b)
-{
-    signalVector *tmp = new signalVector(b->size());
-    tmp->isReal(b->isReal());
-    signalVector::iterator bP = b->begin();
-    signalVector::iterator bPEnd = b->end();
-    signalVector::iterator tmpP = tmp->end()-1;
-    if (!b->isReal()) {
-      while (bP < bPEnd) {
-        *tmpP-- = bP->conj();
-        bP++;
-      }
-    }
-    else {
-      while (bP < bPEnd) {
-        *tmpP-- = bP->real();
-        bP++;
-      }
-    }
-
-    return tmp;
-}
-
 bool vectorSlicer(SoftVector *x)
 {
   SoftVector::iterator xP = x->begin();
@@ -696,7 +558,7 @@ static signalVector *rotateBurst(const BitVector &wBurst,
                                  int guardPeriodLength, int sps)
 {
   int burst_len;
-  signalVector *pulse, rotated, *shaped;
+  signalVector *pulse, rotated;
   signalVector::iterator itr;
 
   pulse = GSMPulse1->empty;
@@ -713,11 +575,7 @@ static signalVector *rotateBurst(const BitVector &wBurst,
   rotated.isReal(false);
 
   /* Dummy filter operation */
-  shaped = convolve(&rotated, pulse, NULL, START_ONLY);
+  return convolve(&rotated, pulse, NULL, START_ONLY);
-  if (!shaped)
-    return NULL;
-
-  return shaped;
 }
 
 static void rotateBurst2(signalVector &burst, double phase)
@@ -737,8 +595,7 @@ static signalVector *modulateBurstLaurent(const BitVector &bits)
 {
   int burst_len, sps = 4;
   float phase;
-  signalVector *c0_pulse, *c1_pulse, *c0_burst;
+  signalVector *c0_pulse, *c1_pulse, *c0_shaped, *c1_shaped;
-  signalVector *c1_burst, *c0_shaped, *c1_shaped;
   signalVector::iterator c0_itr, c1_itr;
 
   c0_pulse = GSMPulse4->c0;
@@ -749,13 +606,12 @@ static signalVector *modulateBurstLaurent(const BitVector &bits)
 
   burst_len = 625;
 
-  c0_burst = new signalVector(burst_len, c0_pulse->size());
+  signalVector c0_burst(burst_len, c0_pulse->size());
-  c0_burst->isReal(true);
+  c0_burst.isReal(true);
-  c0_itr = c0_burst->begin();
+  c0_itr = c0_burst.begin();
 
-  c1_burst = new signalVector(burst_len, c1_pulse->size());
+  signalVector c1_burst(burst_len, c1_pulse->size());
-  c1_burst->isReal(true);
+  c1_itr = c1_burst.begin();
-  c1_itr = c1_burst->begin();
 
   /* Padded differential tail bits */
   *c0_itr = 2.0 * (0x00 & 0x01) - 1.0;
@@ -771,10 +627,10 @@ static signalVector *modulateBurstLaurent(const BitVector &bits)
   *c0_itr = 2.0 * (0x00 & 0x01) - 1.0;
 
   /* Generate C0 phase coefficients */
-  GMSKRotate(*c0_burst, sps);
+  GMSKRotate(c0_burst, sps);
-  c0_burst->isReal(false);
+  c0_burst.isReal(false);
 
-  c0_itr = c0_burst->begin();
+  c0_itr = c0_burst.begin();
   c0_itr += sps * 2;
   c1_itr += sps * 2;
 
@@ -799,8 +655,8 @@ static signalVector *modulateBurstLaurent(const BitVector &bits)
   *c1_itr = *c0_itr * Complex<float>(0, phase);
 
   /* Primary (C0) and secondary (C1) pulse shaping */
-  c0_shaped = convolve(c0_burst, c0_pulse, NULL, START_ONLY);
+  c0_shaped = convolve(&c0_burst, c0_pulse, NULL, START_ONLY);
-  c1_shaped = convolve(c1_burst, c1_pulse, NULL, START_ONLY);
+  c1_shaped = convolve(&c1_burst, c1_pulse, NULL, START_ONLY);
 
   /* Sum shaped outputs into C0 */
   c0_itr = c0_shaped->begin();
@@ -808,10 +664,7 @@ static signalVector *modulateBurstLaurent(const BitVector &bits)
   for (unsigned i = 0; i < c0_shaped->size(); i++ )
     *c0_itr++ += *c1_itr++;
 
-  delete c0_burst;
-  delete c1_burst;
   delete c1_shaped;
-
   return c0_shaped;
 }
 
@@ -885,7 +738,6 @@ static signalVector *mapEdgeSymbols(const BitVector &bits)
 static signalVector *shapeEdgeBurst(const signalVector &symbols)
 {
   size_t nsyms, nsamps = 625, sps = 4;
-  signalVector *burst, *shape;
   signalVector::iterator burst_itr;
 
   nsyms = symbols.size();
@@ -893,10 +745,10 @@ static signalVector *shapeEdgeBurst(const signalVector &symbols)
   if (nsyms * sps > nsamps)
     nsyms = 156;
 
-  burst = new signalVector(nsamps, GSMPulse4->c0->size());
+  signalVector burst(nsamps, GSMPulse4->c0->size());
 
   /* Delay burst by 1 symbol */
-  burst_itr = burst->begin() + sps;
+  burst_itr = burst.begin() + sps;
   for (size_t i = 0; i < nsyms; i++) {
     float phase = i * 3.0f * M_PI / 8.0f;
     Complex<float> rot = Complex<float>(cos(phase), sin(phase));
@@ -906,16 +758,13 @@ static signalVector *shapeEdgeBurst(const signalVector &symbols)
   }
 
   /* Single Gaussian pulse approximation shaping */
-  shape = convolve(burst, GSMPulse4->c0, NULL, START_ONLY);
+  return convolve(&burst, GSMPulse4->c0, NULL, START_ONLY);
-  delete burst;
-
-  return shape;
 }
 
 /*
  * Generate a random GSM normal burst.
  */
-signalVector *genRandNormalBurst(int tsc, int sps, int tn, PRBS &prbs)
+signalVector *genRandNormalBurst(int tsc, int sps, int tn)
 {
   if ((tsc < 0) || (tsc > 7) || (tn < 0) || (tn > 7))
     return NULL;
@@ -923,40 +772,36 @@ signalVector *genRandNormalBurst(int tsc, int sps, int tn, PRBS &prbs)
     return NULL;
 
   int i = 0;
-  BitVector *bits = new BitVector(148);
+  BitVector bits(148);
-  signalVector *burst;
 
   /* Tail bits */
   for (; i < 3; i++)
-    (*bits)[i] = 0;
+    bits[i] = 0;
 
   /* Random bits */
   for (; i < 60; i++)
-    (*bits)[i] = prbs.generateBit();
+    bits[i] = rand() % 2;
 
   /* Stealing bit */
-  (*bits)[i++] = 0;
+  bits[i++] = 0;
 
   /* Training sequence */
   for (int n = 0; i < 87; i++, n++)
-    (*bits)[i] = gTrainingSequence[tsc][n];
+    bits[i] = gTrainingSequence[tsc][n];
 
   /* Stealing bit */
-  (*bits)[i++] = 0;
+  bits[i++] = 0;
 
   /* Random bits */
   for (; i < 145; i++)
-    (*bits)[i] = prbs.generateBit();
+    bits[i] = rand() % 2;
 
   /* Tail bits */
   for (; i < 148; i++)
-    (*bits)[i] = 0;
+    bits[i] = 0;
 
   int guard = 8 + !(tn % 4);
-  burst = modulateBurst(*bits, guard, sps);
+  return modulateBurst(bits, guard, sps);
-  delete bits;
-
-  return burst;
 }
 
 /*
@@ -972,30 +817,26 @@ signalVector *genRandAccessBurst(int delay, int sps, int tn)
     return NULL;
 
   int i = 0;
-  BitVector *bits = new BitVector(88+delay);
+  BitVector bits(88 + delay);
-  signalVector *burst;
 
   /* delay */
   for (; i < delay; i++)
-    (*bits)[i] = 0;
+    bits[i] = 0;
 
   /* head and synch bits */
   for (int n = 0; i < 49+delay; i++, n++)
-    (*bits)[i] = gRACHBurst[n];
+    bits[i] = gRACHBurst[n];
 
   /* Random bits */
   for (; i < 85+delay; i++)
-    (*bits)[i] = rand() % 2;
+    bits[i] = rand() % 2;
 
   /* Tail bits */
   for (; i < 88+delay; i++)
-    (*bits)[i] = 0;
+    bits[i] = 0;
 
   int guard = 68-delay + !(tn % 4);
-  burst = modulateBurst(*bits, guard, sps);
+  return modulateBurst(bits, guard, sps);
-  delete bits;
-
-  return burst;
 }
 
 signalVector *generateEmptyBurst(int sps, int tn)
@@ -1032,17 +873,17 @@ signalVector *generateEdgeBurst(int tsc)
   if ((tsc < 0) || (tsc > 7))
     return NULL;
 
-  signalVector *shape, *burst = new signalVector(148);
+  signalVector burst(148);
   const BitVector *midamble = &gEdgeTrainingSequence[tsc];
 
   /* Tail */
   int n, i = 0;
   for (; i < tail; i++)
-    (*burst)[i] = psk8_table[7];
+    burst[i] = psk8_table[7];
 
   /* Body */
   for (; i < tail + data; i++)
-    (*burst)[i] = psk8_table[rand() % 8];
+    burst[i] = psk8_table[rand() % 8];
 
   /* TSC */
   for (n = 0; i < tail + data + train; i++, n++) {
@@ -1050,21 +891,18 @@ signalVector *generateEdgeBurst(int tsc)
                      (((unsigned) (*midamble)[3 * n + 1] & 0x01) << 1) |
                      (((unsigned) (*midamble)[3 * n + 2] & 0x01) << 2);
 
-    (*burst)[i] = psk8_table[index];
+    burst[i] = psk8_table[index];
   }
 
   /* Body */
   for (; i < tail + data + train + data; i++)
-    (*burst)[i] = psk8_table[rand() % 8];
+    burst[i] = psk8_table[rand() % 8];
 
   /* Tail */
   for (; i < tail + data + train + data + tail; i++)
-    (*burst)[i] = psk8_table[7];
+    burst[i] = psk8_table[7];
 
-  shape = shapeEdgeBurst(*burst);
+  return shapeEdgeBurst(burst);
-  delete burst;
-
-  return shape;
 }
 
 /*
@@ -1100,7 +938,7 @@ static signalVector *modulateBurstBasic(const BitVector &bits,
 					int guard_len, int sps)
 {
   int burst_len;
-  signalVector *pulse, *burst, *shaped;
+  signalVector *pulse;
   signalVector::iterator burst_itr;
 
   if (sps == 1)
@@ -1110,9 +948,9 @@ static signalVector *modulateBurstBasic(const BitVector &bits,
 
   burst_len = sps * (bits.size() + guard_len);
 
-  burst = new signalVector(burst_len, pulse->size());
+  signalVector burst(burst_len, pulse->size());
-  burst->isReal(true);
+  burst.isReal(true);
-  burst_itr = burst->begin();
+  burst_itr = burst.begin();
 
   /* Raw bits are not differentially encoded */
   for (unsigned i = 0; i < bits.size(); i++) {
@@ -1120,15 +958,11 @@ static signalVector *modulateBurstBasic(const BitVector &bits,
     burst_itr += sps;
   }
 
-  GMSKRotate(*burst, sps);
+  GMSKRotate(burst, sps);
-  burst->isReal(false);
+  burst.isReal(false);
 
   /* Single Gaussian pulse approximation shaping */
-  shaped = convolve(burst, pulse, NULL, START_ONLY);
+  return convolve(&burst, pulse, NULL, START_ONLY);
-
-  delete burst;
-
-  return shaped;
 }
 
 /* Assume input bits are not differentially encoded */
@@ -1145,20 +979,14 @@ signalVector *modulateBurst(const BitVector &wBurst, int guardPeriodLength,
 
 static void generateSincTable()
 {
-  float x;
-
   for (int i = 0; i < TABLESIZE; i++) {
-    x = (float) i / TABLESIZE * 8 * M_PI;
+    auto x = (double) i / TABLESIZE * 8 * M_PI;
-    if (fabs(x) < 0.01) {
+    auto y = sin(x) / x;
-      sincTable[i] = 1.0f;
+    sincTable[i] = std::isnan(y) ? 1.0 : y;
-      continue;
-    }
-
-    sincTable[i] = sinf(x) / x;
   }
 }
 
-float sinc(float x)
+static float sinc(float x)
 {
   if (fabs(x) >= 8 * M_PI)
     return 0.0;
@@ -1173,7 +1001,7 @@ float sinc(float x)
  * sinc function generator. The number of filters generated is specified
  * by the DELAYFILTS value.
  */
-void generateDelayFilters()
+static void generateDelayFilters()
 {
   int h_len = 20;
   complex *data;
@@ -1268,31 +1096,8 @@ signalVector *delayVector(const signalVector *in, signalVector *out, float delay
   return out;
 }
 
-signalVector *gaussianNoise(int length, 
+static complex interpolatePoint(const signalVector &inSig, float ix)
-			    float variance, 
-			    complex mean)
 {
-
-  signalVector *noise = new signalVector(length);
-  signalVector::iterator nPtr = noise->begin();
-  float stddev = sqrtf(variance);
-  while (nPtr < noise->end()) {
-    float u1 = (float) rand()/ (float) RAND_MAX;
-    while (u1==0.0)
-      u1 = (float) rand()/ (float) RAND_MAX;
-    float u2 = (float) rand()/ (float) RAND_MAX;
-    float arg = 2.0*M_PI*u2;
-    *nPtr = mean + stddev*complex(cos(arg),sin(arg))*sqrtf(-2.0*log(u1));
-    nPtr++;
-  }
-
-  return noise;
-}
-
-complex interpolatePoint(const signalVector &inSig,
-			 float ix)
-{
-  
   int start = (int) (floor(ix) - 10);
   if (start < 0) start = 0;
   int end = (int) (floor(ix) + 11);
@@ -1332,12 +1137,9 @@ static complex fastPeakDetect(const signalVector &rxBurst, float *index)
   return amp;
 }
 
-complex peakDetect(const signalVector &rxBurst,
+static complex peakDetect(const signalVector &rxBurst,
-		   float *peakIndex,
+                          float *peakIndex, float *avgPwr)
-		   float *avgPwr) 
 {
-  
-
   complex maxVal = 0.0;
   float maxIndex = -1;
   float sumPower = 0.0;
@@ -1410,7 +1212,7 @@ void scaleVector(signalVector &x,
 }
 
 /** in-place conjugation */
-void conjugateVector(signalVector &x)
+static void conjugateVector(signalVector &x)
 {
   if (x.isReal()) return;
   signalVector::iterator xP = x.begin();
@@ -1421,37 +1223,6 @@ void conjugateVector(signalVector &x)
   }
 }
 
-
-// in-place addition!!
-bool addVector(signalVector &x,
-	       signalVector &y)
-{
-  signalVector::iterator xP = x.begin();
-  signalVector::iterator yP = y.begin();
-  signalVector::iterator xPEnd = x.end();
-  signalVector::iterator yPEnd = y.end();
-  while ((xP < xPEnd) && (yP < yPEnd)) {
-    *xP = *xP + *yP;
-    xP++; yP++;
-  }
-  return true;
-}
-
-// in-place multiplication!!
-bool multVector(signalVector &x,
-                 signalVector &y)
-{
-  signalVector::iterator xP = x.begin();
-  signalVector::iterator yP = y.begin();
-  signalVector::iterator xPEnd = x.end();
-  signalVector::iterator yPEnd = y.end();
-  while ((xP < xPEnd) && (yP < yPEnd)) {
-    *xP = (*xP) * (*yP);
-    xP++; yP++;
-  }
-  return true;
-}
-
 static bool generateMidamble(int sps, int tsc)
 {
   bool status = true;
@@ -1528,7 +1299,7 @@ release:
   return status;
 }
 
-CorrelationSequence *generateEdgeMidamble(int tsc)
+static CorrelationSequence *generateEdgeMidamble(int tsc)
 {
   complex *data = NULL;
   signalVector *midamble = NULL, *_midamble = NULL;
@@ -1682,6 +1453,21 @@ float energyDetect(const signalVector &rxBurst, unsigned windowLength)
   return energy/windowLength;
 }
 
+static signalVector *downsampleBurst(const signalVector &burst)
+{
+  signalVector in(DOWNSAMPLE_IN_LEN, dnsampler->len());
+  signalVector *out = new signalVector(DOWNSAMPLE_OUT_LEN);
+  memcpy(in.begin(), burst.begin(), DOWNSAMPLE_IN_LEN * 2 * sizeof(float));
+
+  if (dnsampler->rotate((float *) in.begin(), DOWNSAMPLE_IN_LEN,
+                        (float *) out->begin(), DOWNSAMPLE_OUT_LEN) < 0) {
+    delete out;
+    out = NULL;
+  }
+
+  return out;
+};
+
 /*
  * Detect a burst based on correlation and peak-to-average ratio
  *
@@ -1771,7 +1557,6 @@ static int detectGeneralBurst(const signalVector &rxBurst,
 {
   int rc, start, len;
   bool clipping = false;
-  signalVector *corr;
 
   if ((sps != 1) && (sps != 4))
     return -SIGERR_UNSUPPORTED;
@@ -1787,12 +1572,10 @@ static int detectGeneralBurst(const signalVector &rxBurst,
 
   start = target - head - 1;
   len = head + tail;
-  corr = new signalVector(len);
+  signalVector corr(len);
 
-  rc = detectBurst(rxBurst, *corr, sync,
+  rc = detectBurst(rxBurst, corr, sync,
                    thresh, sps, &amp, &toa, start, len);
-  delete corr;
-
   if (rc < 0) {
     return -SIGERR_INTERNAL;
   } else if (!rc) {
@@ -1816,12 +1599,8 @@ static int detectGeneralBurst(const signalVector &rxBurst,
  *   head: Search 8 symbols before target
  *   tail: Search 8 symbols + maximum expected delay
  */
-int detectRACHBurst(const signalVector &burst,
+static int detectRACHBurst(const signalVector &burst, float threshold, int sps,
-            float threshold,
+                           complex &amplitude, float &toa, unsigned max_toa)
-            int sps,
-            complex &amplitude,
-            float &toa,
-            unsigned max_toa)
 {
   int rc, target, head, tail;
   CorrelationSequence *sync;
@@ -1845,7 +1624,7 @@ int detectRACHBurst(const signalVector &burst,
  *   head: Search 6 symbols before target
  *   tail: Search 6 symbols + maximum expected delay
  */
-int analyzeTrafficBurst(const signalVector &burst, unsigned tsc, float threshold,
+static int analyzeTrafficBurst(const signalVector &burst, unsigned tsc, float threshold,
                                int sps, complex &amplitude, float &toa, unsigned max_toa)
 {
   int rc, target, head, tail;
@@ -1855,8 +1634,8 @@ int analyzeTrafficBurst(const signalVector &burst, unsigned tsc, float threshold
     return -SIGERR_UNSUPPORTED;
 
   target = 3 + 58 + 16 + 5;
-  head = 3;
+  head = 6;
-  tail = 3 + max_toa;
+  tail = 6 + max_toa;
   sync = gMidambles[tsc];
 
   rc = detectGeneralBurst(burst, threshold, sps, amplitude, toa,
@@ -1864,7 +1643,7 @@ int analyzeTrafficBurst(const signalVector &burst, unsigned tsc, float threshold
   return rc;
 }
 
-int detectEdgeBurst(const signalVector &burst, unsigned tsc, float threshold,
+static int detectEdgeBurst(const signalVector &burst, unsigned tsc, float threshold,
                            int sps, complex &amplitude, float &toa, unsigned max_toa)
 {
   int rc, target, head, tail;
@@ -1915,41 +1694,6 @@ int detectAnyBurst(const signalVector &burst, unsigned tsc, float threshold,
   return rc;
 }
 
-signalVector *downsampleBurst(const signalVector &burst)
-{
-  signalVector *in, *out;
-
-  in = new signalVector(DOWNSAMPLE_IN_LEN, dnsampler->len());
-  out = new signalVector(DOWNSAMPLE_OUT_LEN);
-  memcpy(in->begin(), burst.begin(), DOWNSAMPLE_IN_LEN * 2 * sizeof(float));
-
-  if (dnsampler->rotate((float *) in->begin(), DOWNSAMPLE_IN_LEN,
-                        (float *) out->begin(), DOWNSAMPLE_OUT_LEN) < 0) {
-    delete out;
-    out = NULL;
-  }
-
-  delete in;
-  return out;
-};
-
-signalVector *decimateVector(signalVector &wVector, size_t factor)
-{
-  signalVector *dec;
-
-  if (factor <= 1)
-    return NULL;
-
-  dec = new signalVector(wVector.size() / factor);
-  dec->isReal(wVector.isReal());
-
-  signalVector::iterator itr = dec->begin();
-  for (size_t i = 0; i < wVector.size(); i += factor)
-    *itr++ = wVector[i];
-
-  return dec;
-}
-
 /*
  * Soft 8-PSK decoding using Manhattan distance metric
  */
@@ -2046,8 +1790,8 @@ static signalVector *demodCommon(const signalVector &burst, int sps,
  * 4 SPS (if activated) to minimize distortion through the fractional
  * delay filters. Symbol rotation and after always operates at 1 SPS.
  */
-SoftVector *demodGmskBurst(const signalVector &rxBurst, int sps,
+static SoftVector *demodGmskBurst(const signalVector &rxBurst,
-                           complex channel, float TOA)
+                                  int sps, complex channel, float TOA)
 {
   SoftVector *bits;
   signalVector *dec;
@@ -2075,8 +1819,8 @@ SoftVector *demodGmskBurst(const signalVector &rxBurst, int sps,
  * through the fractional delay filters at 1 SPS renders signal
  * nearly unrecoverable.
  */
-SoftVector *demodEdgeBurst(const signalVector &burst, int sps,
+static SoftVector *demodEdgeBurst(const signalVector &burst,
-                           complex chan, float toa)
+                                  int sps, complex chan, float toa)
 {
   SoftVector *bits;
   signalVector *dec, *rot, *eq;
@@ -2110,7 +1854,6 @@ SoftVector *demodAnyBurst(const signalVector &burst, int sps, complex amp,
 
 bool sigProcLibSetup()
 {
-  initTrigTables();
   generateSincTable();
   initGMSKRotationTables();
 
@@ -2137,26 +1880,3 @@ fail:
   sigProcLibDestroy();
   return false;
 }
-
-std::string corrTypeToString(CorrType corr) {
-  switch (corr) {
-  case OFF:
-    return "OFF";
-  case TSC:
-    return "TSC";
-  case RACH:
-    return "RACH";
-  case EDGE:
-    return "EDGE";
-  case IDLE:
-    return "IDLE";
-  default:
-    return "unknown";
-  }
-}
-
-std::ostream& operator<<(std::ostream& os, CorrType corr)
-{
-  os << corrTypeToString(corr);
-  return os;
-}

Transceiver52M/sigProcLib.h

@@ -18,7 +18,6 @@
 #include "Vector.h"
 #include "Complex.h"
 #include "BitVector.h"
-#include "PRBS.h"
 #include "signalVector.h"
 
 /* Burst lengths */
@@ -26,14 +25,6 @@
 #define EDGE_BURST_NBITS      444
 #define EDGE_BURST_NSYMS      (EDGE_BURST_NBITS / 3)
 
-/** Convolution type indicator */
-enum ConvType {
-  START_ONLY,
-  NO_DELAY,
-  CUSTOM,
-  UNDEFINED,
-};
-
 /** Codes for burst types of received bursts*/
 enum CorrType{
   OFF,         ///< timeslot is off
@@ -42,8 +33,6 @@ enum CorrType{
   EDGE,        ///< timeslot should contain an EDGE burst
   IDLE         ///< timeslot is an idle (or dummy) burst
 };
-std::string corrTypeToString(CorrType corr);
-std::ostream& operator<<(std::ostream& os, CorrType corr);
 
 enum SignalError {
   SIGERR_NONE,
@@ -62,66 +51,12 @@ enum SignalError {
  */
 #define BURST_THRESH    4.0
 
-/** Convert a linear number to a dB value */
-float dB(float x);
-
-/** Convert a dB value into a linear value */
-float dBinv(float x);
-
-/** Compute the energy of a vector */
-float vectorNorm2(const signalVector &x);
-
-/** Compute the average power of a vector */
-float vectorPower(const signalVector &x);
-
 /** Setup the signal processing library */
 bool sigProcLibSetup();
 
 /** Destroy the signal processing library */
 void sigProcLibDestroy(void);
 
-/**
-        Convolve two vectors.
-        @param a,b The vectors to be convolved.
-        @param c, A preallocated vector to hold the convolution result.
-        @param spanType The type/span of the convolution.
-        @return The convolution result or NULL on error.
-*/
-signalVector *convolve(const signalVector *a, const signalVector *b,
-                       signalVector *c, ConvType spanType,
-                       size_t start = 0, size_t len = 0,
-                       size_t step = 1, int offset = 0);
-
-/**
-        Frequency shift a vector.
-        @param y The frequency shifted vector.
-        @param x The vector to-be-shifted.
-        @param freq The digital frequency shift
-        @param startPhase The starting phase of the oscillator
-        @param finalPhase The final phase of the oscillator
-        @return The frequency shifted vector.
-*/
-signalVector* frequencyShift(signalVector *y,
-                             signalVector *x,
-                             float freq = 0.0,
-                             float startPhase = 0.0,
-                             float *finalPhase=NULL);
-
-/**
-        Correlate two vectors.
-        @param a,b The vectors to be correlated.
-        @param c, A preallocated vector to hold the correlation result.
-        @param spanType The type/span of the correlation.
-        @return The correlation result.
-*/
-signalVector* correlate(signalVector *a,
-                        signalVector *b,
-                        signalVector *c,
-                        ConvType spanType,
-                        bool bReversedConjugated = false,
-                        unsigned startIx = 0,
-                        unsigned len = 0);
-
 /** Operate soft slicer on a soft-bit vector */
 bool vectorSlicer(SoftVector *x);
 
@@ -141,7 +76,7 @@ signalVector *generateEdgeBurst(int tsc);
 signalVector *generateEmptyBurst(int sps, int tn);
 
 /** Generate a normal GSM burst with random payload - 4 or 1 SPS */
-signalVector *genRandNormalBurst(int tsc, int sps, int tn, PRBS &prbs);
+signalVector *genRandNormalBurst(int tsc, int sps, int tn);
 
 /** Generate an access GSM burst with random payload - 4 or 1 SPS */
 signalVector *genRandAccessBurst(int delay, int sps, int tn);
@@ -149,45 +84,6 @@ signalVector *genRandAccessBurst(int delay, int sps, int tn);
 /** Generate a dummy GSM burst - 4 or 1 SPS */
 signalVector *generateDummyBurst(int sps, int tn);
 
-/** Sinc function */
-float sinc(float x);
-
-/** Delay a vector */
-signalVector *delayVector(const signalVector *in, signalVector *out, float delay);
-
-/** Add two vectors in-place */
-bool addVector(signalVector &x,
-               signalVector &y);
-
-/** Multiply two vectors in-place*/
-bool multVector(signalVector &x,
-                signalVector &y);
-
-/** Generate a vector of gaussian noise */
-signalVector *gaussianNoise(int length,
-                            float variance = 1.0,
-                            complex mean = complex(0.0));
-
-/**
-        Given a non-integer index, interpolate a sample.
-        @param inSig The signal from which to interpolate.
-        @param ix The index.
-        @return The interpolated signal value.
-*/
-complex interpolatePoint(const signalVector &inSig,
-                         float ix);
-
-/**
-        Given a correlator output, locate the correlation peak.
-        @param rxBurst The correlator result.
-        @param peakIndex Pointer to value to receive interpolated peak index.
-        @param avgPower Power to value to receive mean power.
-        @return Peak value.
-*/
-complex peakDetect(const signalVector &rxBurst,
-                   float *peakIndex,
-                   float *avgPwr);
-
 /**
         Apply a scalar to a vector.
         @param x The vector of interest.
@@ -204,68 +100,6 @@ void scaleVector(signalVector &x,
 */
 float energyDetect(const signalVector &rxBurst,
                    unsigned windowLength);
-
-/**
-        RACH aka Access Burst correlator/detector.
-        @param burst The received GSM burst of interest.
-        @param threshold The threshold that the received burst's post-correlator SNR is compared against to determine validity.
-        @param sps The number of samples per GSM symbol.
-        @param amplitude The estimated amplitude of received RACH burst.
-        @param toa The estimate time-of-arrival of received RACH burst.
-        @param max_toa The maximum expected time-of-arrival
-        @return 1 if threshold value is reached,
-                negative value (-SignalError) on error,
-                zero (SIGERR_NONE) if no burst is detected
-*/
-int detectRACHBurst(const signalVector &burst,
-                    float threshold,
-                    int sps,
-                    complex &amplitude,
-                    float &toa,
-                    unsigned max_toa);
-
-/**
-        GMSK Normal Burst correlator/detector.
-        @param rxBurst The received GSM burst of interest.
-        @param tsc Midamble type (0..7) also known as TSC
-        @param threshold The threshold that the received burst's post-correlator SNR is compared against to determine validity.
-        @param sps The number of samples per GSM symbol.
-        @param amplitude The estimated amplitude of received TSC burst.
-        @param toa The estimate time-of-arrival of received TSC burst.
-        @param max_toa The maximum expected time-of-arrival
-        @return 1 if threshold value is reached,
-                negative value (-SignalError) on error,
-                zero (SIGERR_NONE) if no burst is detected
-*/
-int analyzeTrafficBurst(const signalVector &burst,
-                        unsigned tsc,
-                        float threshold,
-                        int sps,
-                        complex &amplitude,
-                        float &toa,
-                        unsigned max_toa);
-
-/**
-        EDGE/8-PSK Normal Burst correlator/detector
-        @param burst The received GSM burst of interest
-        @param tsc Midamble type (0..7) also known as TSC
-        @param threshold The threshold that the received burst's post-correlator SNR is compared against to determine validity.
-        @param sps The number of samples per GSM symbol.
-        @param amplitude The estimated amplitude of received TSC burst.
-        @param toa The estimate time-of-arrival of received TSC burst.
-        @param max_toa The maximum expected time-of-arrival
-        @return 1 if threshold value is reached,
-                negative value (-SignalError) on error,
-                zero (SIGERR_NONE) if no burst is detected
-*/
-int detectEdgeBurst(const signalVector &burst,
-                    unsigned tsc,
-                    float threshold,
-                    int sps,
-                    complex &amplitude,
-                    float &toa,
-                    unsigned max_toa);
-
 /**
         8-PSK/GMSK/RACH burst detector
         @param burst The received GSM burst of interest
@@ -288,44 +122,6 @@ int detectAnyBurst(const signalVector &burst,
                    float &toa,
                    unsigned max_toa);
 
-/**
-        Downsample 4 SPS to 1 SPS using a polyphase filterbank
-        @param burst Input burst of at least 624 symbols
-        @return Decimated signal vector of 156 symbols
-*/
-signalVector *downsampleBurst(const signalVector &burst);
-
-/**
-        Decimate a vector.
-        @param wVector The vector of interest.
-        @param factor Decimation factor.
-        @return The decimated signal vector.
-*/
-signalVector *decimateVector(signalVector &wVector, size_t factor);
-
-/**
-        Demodulates a GMSK burst using a soft-slicer.
-        @param rxBurst The burst to be demodulated.
-        @param gsmPulse The GSM pulse.
-        @param sps The number of samples per GSM symbol.
-        @param channel The amplitude estimate of the received burst.
-        @param TOA The time-of-arrival of the received burst.
-        @return The demodulated bit sequence.
-*/
-SoftVector *demodGmskBurst(const signalVector &rxBurst, int sps,
-                           complex channel, float TOA);
-
-/**
-        Demodulate 8-PSK EDGE burst with soft symbol ooutput
-        @param rxBurst The burst to be demodulated.
-        @param sps The number of samples per GSM symbol.
-        @param channel The amplitude estimate of the received burst.
-        @param TOA The time-of-arrival of the received burst.
-        @return The demodulated bit sequence.
-*/
-SoftVector *demodEdgeBurst(const signalVector &rxBurst, int sps,
-                           complex channel, float TOA);
-
 /** Demodulate burst basde on type and output soft bits */
 SoftVector *demodAnyBurst(const signalVector &burst, int sps,
                           complex amp, float toa, CorrType type);

Transceiver52M/x86/convert.c

@@ -46,6 +46,7 @@ void convert_init(void)
 	c.convert_si16_ps_16n = base_convert_short_float;
 	c.convert_si16_ps = base_convert_short_float;
 
+#ifdef HAVE___BUILTIN_CPU_SUPPORTS
 #ifdef HAVE_SSE4_1
 	if (__builtin_cpu_supports("sse4.1")) {
 		c.convert_si16_ps_16n = &_sse_convert_si16_ps_16n;
@@ -60,6 +61,7 @@ void convert_init(void)
 		c.convert_scale_ps_si16 = _sse_convert_scale_ps_si16;
 	}
 #endif
+#endif
 }
 
 void convert_float_short(short *out, const float *in, float scale, int len)

Transceiver52M/x86/convolve.c

@@ -82,7 +82,7 @@ void convolve_init(void)
 	c.conv_real4n = (void *)_base_convolve_real;
 	c.conv_real = (void *)_base_convolve_real;
 
-#ifdef HAVE_SSE3
+#if defined(HAVE_SSE3) && defined(HAVE___BUILTIN_CPU_SUPPORTS)
 	if (__builtin_cpu_supports("sse3")) {
 		c.conv_cmplx_4n = sse_conv_cmplx_4n;
 		c.conv_cmplx_8n = sse_conv_cmplx_8n;

config/ax_cxx_compile_stdcxx.m4

@@ -0,0 +1,982 @@
+# ===========================================================================
+#  https://www.gnu.org/software/autoconf-archive/ax_cxx_compile_stdcxx.html
+# ===========================================================================
+#
+# SYNOPSIS
+#
+#   AX_CXX_COMPILE_STDCXX(VERSION, [ext|noext], [mandatory|optional])
+#
+# DESCRIPTION
+#
+#   Check for baseline language coverage in the compiler for the specified
+#   version of the C++ standard.  If necessary, add switches to CXX and
+#   CXXCPP to enable support.  VERSION may be '11' (for the C++11 standard)
+#   or '14' (for the C++14 standard).
+#
+#   The second argument, if specified, indicates whether you insist on an
+#   extended mode (e.g. -std=gnu++11) or a strict conformance mode (e.g.
+#   -std=c++11).  If neither is specified, you get whatever works, with
+#   preference for an extended mode.
+#
+#   The third argument, if specified 'mandatory' or if left unspecified,
+#   indicates that baseline support for the specified C++ standard is
+#   required and that the macro should error out if no mode with that
+#   support is found.  If specified 'optional', then configuration proceeds
+#   regardless, after defining HAVE_CXX${VERSION} if and only if a
+#   supporting mode is found.
+#
+# LICENSE
+#
+#   Copyright (c) 2008 Benjamin Kosnik <bkoz@redhat.com>
+#   Copyright (c) 2012 Zack Weinberg <zackw@panix.com>
+#   Copyright (c) 2013 Roy Stogner <roystgnr@ices.utexas.edu>
+#   Copyright (c) 2014, 2015 Google Inc.; contributed by Alexey Sokolov <sokolov@google.com>
+#   Copyright (c) 2015 Paul Norman <penorman@mac.com>
+#   Copyright (c) 2015 Moritz Klammler <moritz@klammler.eu>
+#   Copyright (c) 2016 Krzesimir Nowak <qdlacz@gmail.com>
+#
+#   Copying and distribution of this file, with or without modification, are
+#   permitted in any medium without royalty provided the copyright notice
+#   and this notice are preserved.  This file is offered as-is, without any
+#   warranty.
+
+#serial 7
+
+dnl  This macro is based on the code from the AX_CXX_COMPILE_STDCXX_11 macro
+dnl  (serial version number 13).
+
+AX_REQUIRE_DEFINED([AC_MSG_WARN])
+AC_DEFUN([AX_CXX_COMPILE_STDCXX], [dnl
+  m4_if([$1], [11], [ax_cxx_compile_alternatives="11 0x"],
+        [$1], [14], [ax_cxx_compile_alternatives="14 1y"],
+        [$1], [17], [ax_cxx_compile_alternatives="17 1z"],
+        [m4_fatal([invalid first argument `$1' to AX_CXX_COMPILE_STDCXX])])dnl
+  m4_if([$2], [], [],
+        [$2], [ext], [],
+        [$2], [noext], [],
+        [m4_fatal([invalid second argument `$2' to AX_CXX_COMPILE_STDCXX])])dnl
+  m4_if([$3], [], [ax_cxx_compile_cxx$1_required=true],
+        [$3], [mandatory], [ax_cxx_compile_cxx$1_required=true],
+        [$3], [optional], [ax_cxx_compile_cxx$1_required=false],
+        [m4_fatal([invalid third argument `$3' to AX_CXX_COMPILE_STDCXX])])
+  AC_LANG_PUSH([C++])dnl
+  ac_success=no
+  AC_CACHE_CHECK(whether $CXX supports C++$1 features by default,
+  ax_cv_cxx_compile_cxx$1,
+  [AC_COMPILE_IFELSE([AC_LANG_SOURCE([_AX_CXX_COMPILE_STDCXX_testbody_$1])],
+    [ax_cv_cxx_compile_cxx$1=yes],
+    [ax_cv_cxx_compile_cxx$1=no])])
+  if test x$ax_cv_cxx_compile_cxx$1 = xyes; then
+    ac_success=yes
+  fi
+
+  m4_if([$2], [noext], [], [dnl
+  if test x$ac_success = xno; then
+    for alternative in ${ax_cxx_compile_alternatives}; do
+      switch="-std=gnu++${alternative}"
+      cachevar=AS_TR_SH([ax_cv_cxx_compile_cxx$1_$switch])
+      AC_CACHE_CHECK(whether $CXX supports C++$1 features with $switch,
+                     $cachevar,
+        [ac_save_CXX="$CXX"
+         CXX="$CXX $switch"
+         AC_COMPILE_IFELSE([AC_LANG_SOURCE([_AX_CXX_COMPILE_STDCXX_testbody_$1])],
+          [eval $cachevar=yes],
+          [eval $cachevar=no])
+         CXX="$ac_save_CXX"])
+      if eval test x\$$cachevar = xyes; then
+        CXX="$CXX $switch"
+        if test -n "$CXXCPP" ; then
+          CXXCPP="$CXXCPP $switch"
+        fi
+        ac_success=yes
+        break
+      fi
+    done
+  fi])
+
+  m4_if([$2], [ext], [], [dnl
+  if test x$ac_success = xno; then
+    dnl HP's aCC needs +std=c++11 according to:
+    dnl http://h21007.www2.hp.com/portal/download/files/unprot/aCxx/PDF_Release_Notes/769149-001.pdf
+    dnl Cray's crayCC needs "-h std=c++11"
+    for alternative in ${ax_cxx_compile_alternatives}; do
+      for switch in -std=c++${alternative} +std=c++${alternative} "-h std=c++${alternative}"; do
+        cachevar=AS_TR_SH([ax_cv_cxx_compile_cxx$1_$switch])
+        AC_CACHE_CHECK(whether $CXX supports C++$1 features with $switch,
+                       $cachevar,
+          [ac_save_CXX="$CXX"
+           CXX="$CXX $switch"
+           AC_COMPILE_IFELSE([AC_LANG_SOURCE([_AX_CXX_COMPILE_STDCXX_testbody_$1])],
+            [eval $cachevar=yes],
+            [eval $cachevar=no])
+           CXX="$ac_save_CXX"])
+        if eval test x\$$cachevar = xyes; then
+          CXX="$CXX $switch"
+          if test -n "$CXXCPP" ; then
+            CXXCPP="$CXXCPP $switch"
+          fi
+          ac_success=yes
+          break
+        fi
+      done
+      if test x$ac_success = xyes; then
+        break
+      fi
+    done
+  fi])
+  AC_LANG_POP([C++])
+  if test x$ax_cxx_compile_cxx$1_required = xtrue; then
+    if test x$ac_success = xno; then
+      AC_MSG_ERROR([*** A compiler with support for C++$1 language features is required.])
+    fi
+  fi
+  if test x$ac_success = xno; then
+    HAVE_CXX$1=0
+    AC_MSG_NOTICE([No compiler with C++$1 support was found])
+  else
+    HAVE_CXX$1=1
+    AC_DEFINE(HAVE_CXX$1,1,
+              [define if the compiler supports basic C++$1 syntax])
+  fi
+  AC_SUBST(HAVE_CXX$1)
+  m4_if([$1], [17], [AC_MSG_WARN([C++17 is not yet standardized, so the checks may change in incompatible ways anytime])])
+])
+
+
+dnl  Test body for checking C++11 support
+
+m4_define([_AX_CXX_COMPILE_STDCXX_testbody_11],
+  _AX_CXX_COMPILE_STDCXX_testbody_new_in_11
+)
+
+
+dnl  Test body for checking C++14 support
+
+m4_define([_AX_CXX_COMPILE_STDCXX_testbody_14],
+  _AX_CXX_COMPILE_STDCXX_testbody_new_in_11
+  _AX_CXX_COMPILE_STDCXX_testbody_new_in_14
+)
+
+m4_define([_AX_CXX_COMPILE_STDCXX_testbody_17],
+  _AX_CXX_COMPILE_STDCXX_testbody_new_in_11
+  _AX_CXX_COMPILE_STDCXX_testbody_new_in_14
+  _AX_CXX_COMPILE_STDCXX_testbody_new_in_17
+)
+
+dnl  Tests for new features in C++11
+
+m4_define([_AX_CXX_COMPILE_STDCXX_testbody_new_in_11], [[
+
+// If the compiler admits that it is not ready for C++11, why torture it?
+// Hopefully, this will speed up the test.
+
+#ifndef __cplusplus
+
+#error "This is not a C++ compiler"
+
+#elif __cplusplus < 201103L
+
+#error "This is not a C++11 compiler"
+
+#else
+
+namespace cxx11
+{
+
+  namespace test_static_assert
+  {
+
+    template <typename T>
+    struct check
+    {
+      static_assert(sizeof(int) <= sizeof(T), "not big enough");
+    };
+
+  }
+
+  namespace test_final_override
+  {
+
+    struct Base
+    {
+      virtual void f() {}
+    };
+
+    struct Derived : public Base
+    {
+      virtual void f() override {}
+    };
+
+  }
+
+  namespace test_double_right_angle_brackets
+  {
+
+    template < typename T >
+    struct check {};
+
+    typedef check<void> single_type;
+    typedef check<check<void>> double_type;
+    typedef check<check<check<void>>> triple_type;
+    typedef check<check<check<check<void>>>> quadruple_type;
+
+  }
+
+  namespace test_decltype
+  {
+
+    int
+    f()
+    {
+      int a = 1;
+      decltype(a) b = 2;
+      return a + b;
+    }
+
+  }
+
+  namespace test_type_deduction
+  {
+
+    template < typename T1, typename T2 >
+    struct is_same
+    {
+      static const bool value = false;
+    };
+
+    template < typename T >
+    struct is_same<T, T>
+    {
+      static const bool value = true;
+    };
+
+    template < typename T1, typename T2 >
+    auto
+    add(T1 a1, T2 a2) -> decltype(a1 + a2)
+    {
+      return a1 + a2;
+    }
+
+    int
+    test(const int c, volatile int v)
+    {
+      static_assert(is_same<int, decltype(0)>::value == true, "");
+      static_assert(is_same<int, decltype(c)>::value == false, "");
+      static_assert(is_same<int, decltype(v)>::value == false, "");
+      auto ac = c;
+      auto av = v;
+      auto sumi = ac + av + 'x';
+      auto sumf = ac + av + 1.0;
+      static_assert(is_same<int, decltype(ac)>::value == true, "");
+      static_assert(is_same<int, decltype(av)>::value == true, "");
+      static_assert(is_same<int, decltype(sumi)>::value == true, "");
+      static_assert(is_same<int, decltype(sumf)>::value == false, "");
+      static_assert(is_same<int, decltype(add(c, v))>::value == true, "");
+      return (sumf > 0.0) ? sumi : add(c, v);
+    }
+
+  }
+
+  namespace test_noexcept
+  {
+
+    int f() { return 0; }
+    int g() noexcept { return 0; }
+
+    static_assert(noexcept(f()) == false, "");
+    static_assert(noexcept(g()) == true, "");
+
+  }
+
+  namespace test_constexpr
+  {
+
+    template < typename CharT >
+    unsigned long constexpr
+    strlen_c_r(const CharT *const s, const unsigned long acc) noexcept
+    {
+      return *s ? strlen_c_r(s + 1, acc + 1) : acc;
+    }
+
+    template < typename CharT >
+    unsigned long constexpr
+    strlen_c(const CharT *const s) noexcept
+    {
+      return strlen_c_r(s, 0UL);
+    }
+
+    static_assert(strlen_c("") == 0UL, "");
+    static_assert(strlen_c("1") == 1UL, "");
+    static_assert(strlen_c("example") == 7UL, "");
+    static_assert(strlen_c("another\0example") == 7UL, "");
+
+  }
+
+  namespace test_rvalue_references
+  {
+
+    template < int N >
+    struct answer
+    {
+      static constexpr int value = N;
+    };
+
+    answer<1> f(int&)       { return answer<1>(); }
+    answer<2> f(const int&) { return answer<2>(); }
+    answer<3> f(int&&)      { return answer<3>(); }
+
+    void
+    test()
+    {
+      int i = 0;
+      const int c = 0;
+      static_assert(decltype(f(i))::value == 1, "");
+      static_assert(decltype(f(c))::value == 2, "");
+      static_assert(decltype(f(0))::value == 3, "");
+    }
+
+  }
+
+  namespace test_uniform_initialization
+  {
+
+    struct test
+    {
+      static const int zero {};
+      static const int one {1};
+    };
+
+    static_assert(test::zero == 0, "");
+    static_assert(test::one == 1, "");
+
+  }
+
+  namespace test_lambdas
+  {
+
+    void
+    test1()
+    {
+      auto lambda1 = [](){};
+      auto lambda2 = lambda1;
+      lambda1();
+      lambda2();
+    }
+
+    int
+    test2()
+    {
+      auto a = [](int i, int j){ return i + j; }(1, 2);
+      auto b = []() -> int { return '0'; }();
+      auto c = [=](){ return a + b; }();
+      auto d = [&](){ return c; }();
+      auto e = [a, &b](int x) mutable {
+        const auto identity = [](int y){ return y; };
+        for (auto i = 0; i < a; ++i)
+          a += b--;
+        return x + identity(a + b);
+      }(0);
+      return a + b + c + d + e;
+    }
+
+    int
+    test3()
+    {
+      const auto nullary = [](){ return 0; };
+      const auto unary = [](int x){ return x; };
+      using nullary_t = decltype(nullary);
+      using unary_t = decltype(unary);
+      const auto higher1st = [](nullary_t f){ return f(); };
+      const auto higher2nd = [unary](nullary_t f1){
+        return [unary, f1](unary_t f2){ return f2(unary(f1())); };
+      };
+      return higher1st(nullary) + higher2nd(nullary)(unary);
+    }
+
+  }
+
+  namespace test_variadic_templates
+  {
+
+    template <int...>
+    struct sum;
+
+    template <int N0, int... N1toN>
+    struct sum<N0, N1toN...>
+    {
+      static constexpr auto value = N0 + sum<N1toN...>::value;
+    };
+
+    template <>
+    struct sum<>
+    {
+      static constexpr auto value = 0;
+    };
+
+    static_assert(sum<>::value == 0, "");
+    static_assert(sum<1>::value == 1, "");
+    static_assert(sum<23>::value == 23, "");
+    static_assert(sum<1, 2>::value == 3, "");
+    static_assert(sum<5, 5, 11>::value == 21, "");
+    static_assert(sum<2, 3, 5, 7, 11, 13>::value == 41, "");
+
+  }
+
+  // http://stackoverflow.com/questions/13728184/template-aliases-and-sfinae
+  // Clang 3.1 fails with headers of libstd++ 4.8.3 when using std::function
+  // because of this.
+  namespace test_template_alias_sfinae
+  {
+
+    struct foo {};
+
+    template<typename T>
+    using member = typename T::member_type;
+
+    template<typename T>
+    void func(...) {}
+
+    template<typename T>
+    void func(member<T>*) {}
+
+    void test();
+
+    void test() { func<foo>(0); }
+
+  }
+
+}  // namespace cxx11
+
+#endif  // __cplusplus >= 201103L
+
+]])
+
+
+dnl  Tests for new features in C++14
+
+m4_define([_AX_CXX_COMPILE_STDCXX_testbody_new_in_14], [[
+
+// If the compiler admits that it is not ready for C++14, why torture it?
+// Hopefully, this will speed up the test.
+
+#ifndef __cplusplus
+
+#error "This is not a C++ compiler"
+
+#elif __cplusplus < 201402L
+
+#error "This is not a C++14 compiler"
+
+#else
+
+namespace cxx14
+{
+
+  namespace test_polymorphic_lambdas
+  {
+
+    int
+    test()
+    {
+      const auto lambda = [](auto&&... args){
+        const auto istiny = [](auto x){
+          return (sizeof(x) == 1UL) ? 1 : 0;
+        };
+        const int aretiny[] = { istiny(args)... };
+        return aretiny[0];
+      };
+      return lambda(1, 1L, 1.0f, '1');
+    }
+
+  }
+
+  namespace test_binary_literals
+  {
+
+    constexpr auto ivii = 0b0000000000101010;
+    static_assert(ivii == 42, "wrong value");
+
+  }
+
+  namespace test_generalized_constexpr
+  {
+
+    template < typename CharT >
+    constexpr unsigned long
+    strlen_c(const CharT *const s) noexcept
+    {
+      auto length = 0UL;
+      for (auto p = s; *p; ++p)
+        ++length;
+      return length;
+    }
+
+    static_assert(strlen_c("") == 0UL, "");
+    static_assert(strlen_c("x") == 1UL, "");
+    static_assert(strlen_c("test") == 4UL, "");
+    static_assert(strlen_c("another\0test") == 7UL, "");
+
+  }
+
+  namespace test_lambda_init_capture
+  {
+
+    int
+    test()
+    {
+      auto x = 0;
+      const auto lambda1 = [a = x](int b){ return a + b; };
+      const auto lambda2 = [a = lambda1(x)](){ return a; };
+      return lambda2();
+    }
+
+  }
+
+  namespace test_digit_separators
+  {
+
+    constexpr auto ten_million = 100'000'000;
+    static_assert(ten_million == 100000000, "");
+
+  }
+
+  namespace test_return_type_deduction
+  {
+
+    auto f(int& x) { return x; }
+    decltype(auto) g(int& x) { return x; }
+
+    template < typename T1, typename T2 >
+    struct is_same
+    {
+      static constexpr auto value = false;
+    };
+
+    template < typename T >
+    struct is_same<T, T>
+    {
+      static constexpr auto value = true;
+    };
+
+    int
+    test()
+    {
+      auto x = 0;
+      static_assert(is_same<int, decltype(f(x))>::value, "");
+      static_assert(is_same<int&, decltype(g(x))>::value, "");
+      return x;
+    }
+
+  }
+
+}  // namespace cxx14
+
+#endif  // __cplusplus >= 201402L
+
+]])
+
+
+dnl  Tests for new features in C++17
+
+m4_define([_AX_CXX_COMPILE_STDCXX_testbody_new_in_17], [[
+
+// If the compiler admits that it is not ready for C++17, why torture it?
+// Hopefully, this will speed up the test.
+
+#ifndef __cplusplus
+
+#error "This is not a C++ compiler"
+
+#elif __cplusplus <= 201402L
+
+#error "This is not a C++17 compiler"
+
+#else
+
+#if defined(__clang__)
+  #define REALLY_CLANG
+#else
+  #if defined(__GNUC__)
+    #define REALLY_GCC
+  #endif
+#endif
+
+#include <initializer_list>
+#include <utility>
+#include <type_traits>
+
+namespace cxx17
+{
+
+#if !defined(REALLY_CLANG)
+  namespace test_constexpr_lambdas
+  {
+
+    // TODO: test it with clang++ from git
+
+    constexpr int foo = [](){return 42;}();
+
+  }
+#endif // !defined(REALLY_CLANG)
+
+  namespace test::nested_namespace::definitions
+  {
+
+  }
+
+  namespace test_fold_expression
+  {
+
+    template<typename... Args>
+    int multiply(Args... args)
+    {
+      return (args * ... * 1);
+    }
+
+    template<typename... Args>
+    bool all(Args... args)
+    {
+      return (args && ...);
+    }
+
+  }
+
+  namespace test_extended_static_assert
+  {
+
+    static_assert (true);
+
+  }
+
+  namespace test_auto_brace_init_list
+  {
+
+    auto foo = {5};
+    auto bar {5};
+
+    static_assert(std::is_same<std::initializer_list<int>, decltype(foo)>::value);
+    static_assert(std::is_same<int, decltype(bar)>::value);
+  }
+
+  namespace test_typename_in_template_template_parameter
+  {
+
+    template<template<typename> typename X> struct D;
+
+  }
+
+  namespace test_fallthrough_nodiscard_maybe_unused_attributes
+  {
+
+    int f1()
+    {
+      return 42;
+    }
+
+    [[nodiscard]] int f2()
+    {
+      [[maybe_unused]] auto unused = f1();
+
+      switch (f1())
+      {
+      case 17:
+        f1();
+        [[fallthrough]];
+      case 42:
+        f1();
+      }
+      return f1();
+    }
+
+  }
+
+  namespace test_extended_aggregate_initialization
+  {
+
+    struct base1
+    {
+      int b1, b2 = 42;
+    };
+
+    struct base2
+    {
+      base2() {
+        b3 = 42;
+      }
+      int b3;
+    };
+
+    struct derived : base1, base2
+    {
+        int d;
+    };
+
+    derived d1 {{1, 2}, {}, 4};  // full initialization
+    derived d2 {{}, {}, 4};      // value-initialized bases
+
+  }
+
+  namespace test_general_range_based_for_loop
+  {
+
+    struct iter
+    {
+      int i;
+
+      int& operator* ()
+      {
+        return i;
+      }
+
+      const int& operator* () const
+      {
+        return i;
+      }
+
+      iter& operator++()
+      {
+        ++i;
+        return *this;
+      }
+    };
+
+    struct sentinel
+    {
+      int i;
+    };
+
+    bool operator== (const iter& i, const sentinel& s)
+    {
+      return i.i == s.i;
+    }
+
+    bool operator!= (const iter& i, const sentinel& s)
+    {
+      return !(i == s);
+    }
+
+    struct range
+    {
+      iter begin() const
+      {
+        return {0};
+      }
+
+      sentinel end() const
+      {
+        return {5};
+      }
+    };
+
+    void f()
+    {
+      range r {};
+
+      for (auto i : r)
+      {
+        [[maybe_unused]] auto v = i;
+      }
+    }
+
+  }
+
+  namespace test_lambda_capture_asterisk_this_by_value
+  {
+
+    struct t
+    {
+      int i;
+      int foo()
+      {
+        return [*this]()
+        {
+          return i;
+        }();
+      }
+    };
+
+  }
+
+  namespace test_enum_class_construction
+  {
+
+    enum class byte : unsigned char
+    {};
+
+    byte foo {42};
+
+  }
+
+  namespace test_constexpr_if
+  {
+
+    template <bool cond>
+    int f ()
+    {
+      if constexpr(cond)
+      {
+        return 13;
+      }
+      else
+      {
+        return 42;
+      }
+    }
+
+  }
+
+  namespace test_selection_statement_with_initializer
+  {
+
+    int f()
+    {
+      return 13;
+    }
+
+    int f2()
+    {
+      if (auto i = f(); i > 0)
+      {
+        return 3;
+      }
+
+      switch (auto i = f(); i + 4)
+      {
+      case 17:
+        return 2;
+
+      default:
+        return 1;
+      }
+    }
+
+  }
+
+#if !defined(REALLY_CLANG)
+  namespace test_template_argument_deduction_for_class_templates
+  {
+
+    // TODO: test it with clang++ from git
+
+    template <typename T1, typename T2>
+    struct pair
+    {
+      pair (T1 p1, T2 p2)
+        : m1 {p1},
+          m2 {p2}
+      {}
+
+      T1 m1;
+      T2 m2;
+    };
+
+    void f()
+    {
+      [[maybe_unused]] auto p = pair{13, 42u};
+    }
+
+  }
+#endif // !defined(REALLY_CLANG)
+
+  namespace test_non_type_auto_template_parameters
+  {
+
+    template <auto n>
+    struct B
+    {};
+
+    B<5> b1;
+    B<'a'> b2;
+
+  }
+
+#if !defined(REALLY_CLANG)
+  namespace test_structured_bindings
+  {
+
+    // TODO: test it with clang++ from git
+
+    int arr[2] = { 1, 2 };
+    std::pair<int, int> pr = { 1, 2 };
+
+    auto f1() -> int(&)[2]
+    {
+      return arr;
+    }
+
+    auto f2() -> std::pair<int, int>&
+    {
+      return pr;
+    }
+
+    struct S
+    {
+      int x1 : 2;
+      volatile double y1;
+    };
+
+    S f3()
+    {
+      return {};
+    }
+
+    auto [ x1, y1 ] = f1();
+    auto& [ xr1, yr1 ] = f1();
+    auto [ x2, y2 ] = f2();
+    auto& [ xr2, yr2 ] = f2();
+    const auto [ x3, y3 ] = f3();
+
+  }
+#endif // !defined(REALLY_CLANG)
+
+#if !defined(REALLY_CLANG)
+  namespace test_exception_spec_type_system
+  {
+
+    // TODO: test it with clang++ from git
+
+    struct Good {};
+    struct Bad {};
+
+    void g1() noexcept;
+    void g2();
+
+    template<typename T>
+    Bad
+    f(T*, T*);
+
+    template<typename T1, typename T2>
+    Good
+    f(T1*, T2*);
+
+    static_assert (std::is_same_v<Good, decltype(f(g1, g2))>);
+
+  }
+#endif // !defined(REALLY_CLANG)
+
+  namespace test_inline_variables
+  {
+
+    template<class T> void f(T)
+    {}
+
+    template<class T> inline T g(T)
+    {
+      return T{};
+    }
+
+    template<> inline void f<>(int)
+    {}
+
+    template<> int g<>(int)
+    {
+      return 5;
+    }
+
+  }
+
+}  // namespace cxx17
+
+#endif  // __cplusplus <= 201402L
+
+]])

config/ax_cxx_compile_stdcxx_11.m4

@@ -0,0 +1,39 @@
+# =============================================================================
+#  https://www.gnu.org/software/autoconf-archive/ax_cxx_compile_stdcxx_11.html
+# =============================================================================
+#
+# SYNOPSIS
+#
+#   AX_CXX_COMPILE_STDCXX_11([ext|noext], [mandatory|optional])
+#
+# DESCRIPTION
+#
+#   Check for baseline language coverage in the compiler for the C++11
+#   standard; if necessary, add switches to CXX and CXXCPP to enable
+#   support.
+#
+#   This macro is a convenience alias for calling the AX_CXX_COMPILE_STDCXX
+#   macro with the version set to C++11.  The two optional arguments are
+#   forwarded literally as the second and third argument respectively.
+#   Please see the documentation for the AX_CXX_COMPILE_STDCXX macro for
+#   more information.  If you want to use this macro, you also need to
+#   download the ax_cxx_compile_stdcxx.m4 file.
+#
+# LICENSE
+#
+#   Copyright (c) 2008 Benjamin Kosnik <bkoz@redhat.com>
+#   Copyright (c) 2012 Zack Weinberg <zackw@panix.com>
+#   Copyright (c) 2013 Roy Stogner <roystgnr@ices.utexas.edu>
+#   Copyright (c) 2014, 2015 Google Inc.; contributed by Alexey Sokolov <sokolov@google.com>
+#   Copyright (c) 2015 Paul Norman <penorman@mac.com>
+#   Copyright (c) 2015 Moritz Klammler <moritz@klammler.eu>
+#
+#   Copying and distribution of this file, with or without modification, are
+#   permitted in any medium without royalty provided the copyright notice
+#   and this notice are preserved. This file is offered as-is, without any
+#   warranty.
+
+#serial 18
+
+AX_REQUIRE_DEFINED([AX_CXX_COMPILE_STDCXX])
+AC_DEFUN([AX_CXX_COMPILE_STDCXX_11], [AX_CXX_COMPILE_STDCXX([11], [$1], [$2])])

config/ax_gcc_x86_avx_xgetbv.m4

@@ -1,79 +0,0 @@
-# ===========================================================================
-#   http://www.gnu.org/software/autoconf-archive/ax_gcc_x86_avx_xgetbv.html
-# ===========================================================================
-#
-# SYNOPSIS
-#
-#   AX_GCC_X86_AVX_XGETBV
-#
-# DESCRIPTION
-#
-#   On later x86 processors with AVX SIMD support, with gcc or a compiler
-#   that has a compatible syntax for inline assembly instructions, run a
-#   small program that executes the xgetbv instruction with input OP. This
-#   can be used to detect if the OS supports AVX instruction usage.
-#
-#   On output, the values of the eax and edx registers are stored as
-#   hexadecimal strings as "eax:edx" in the cache variable
-#   ax_cv_gcc_x86_avx_xgetbv.
-#
-#   If the xgetbv instruction fails (because you are running a
-#   cross-compiler, or because you are not using gcc, or because you are on
-#   a processor that doesn't have this instruction),
-#   ax_cv_gcc_x86_avx_xgetbv_OP is set to the string "unknown".
-#
-#   This macro mainly exists to be used in AX_EXT.
-#
-# LICENSE
-#
-#   Copyright (c) 2013 Michael Petch <mpetch@capp-sysware.com>
-#
-#   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, either 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 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/>.
-#
-#   As a special exception, the respective Autoconf Macro's copyright owner
-#   gives unlimited permission to copy, distribute and modify the configure
-#   scripts that are the output of Autoconf when processing the Macro. You
-#   need not follow the terms of the GNU General Public License when using
-#   or distributing such scripts, even though portions of the text of the
-#   Macro appear in them. The GNU General Public License (GPL) does govern
-#   all other use of the material that constitutes the Autoconf Macro.
-#
-#   This special exception to the GPL applies to versions of the Autoconf
-#   Macro released by the Autoconf Archive. When you make and distribute a
-#   modified version of the Autoconf Macro, you may extend this special
-#   exception to the GPL to apply to your modified version as well.
-
-#serial 1
-
-AC_DEFUN([AX_GCC_X86_AVX_XGETBV],
-[AC_REQUIRE([AC_PROG_CC])
-AC_LANG_PUSH([C])
-AC_CACHE_CHECK(for x86-AVX xgetbv $1 output, ax_cv_gcc_x86_avx_xgetbv_$1,
- [AC_RUN_IFELSE([AC_LANG_PROGRAM([#include <stdio.h>], [
-     int op = $1, eax, edx;
-     FILE *f;
-      /* Opcodes for xgetbv */
-      __asm__(".byte 0x0f, 0x01, 0xd0"
-        : "=a" (eax), "=d" (edx)
-        : "c" (op));
-     f = fopen("conftest_xgetbv", "w"); if (!f) return 1;
-     fprintf(f, "%x:%x\n", eax, edx);
-     fclose(f);
-     return 0;
-])],
-     [ax_cv_gcc_x86_avx_xgetbv_$1=`cat conftest_xgetbv`; rm -f conftest_xgetbv],
-     [ax_cv_gcc_x86_avx_xgetbv_$1=unknown; rm -f conftest_xgetbv],
-     [ax_cv_gcc_x86_avx_xgetbv_$1=unknown])])
-AC_LANG_POP([C])
-])

config/ax_gcc_x86_cpuid.m4

@@ -1,79 +0,0 @@
-# ===========================================================================
-#     http://www.gnu.org/software/autoconf-archive/ax_gcc_x86_cpuid.html
-# ===========================================================================
-#
-# SYNOPSIS
-#
-#   AX_GCC_X86_CPUID(OP)
-#
-# DESCRIPTION
-#
-#   On Pentium and later x86 processors, with gcc or a compiler that has a
-#   compatible syntax for inline assembly instructions, run a small program
-#   that executes the cpuid instruction with input OP. This can be used to
-#   detect the CPU type.
-#
-#   On output, the values of the eax, ebx, ecx, and edx registers are stored
-#   as hexadecimal strings as "eax:ebx:ecx:edx" in the cache variable
-#   ax_cv_gcc_x86_cpuid_OP.
-#
-#   If the cpuid instruction fails (because you are running a
-#   cross-compiler, or because you are not using gcc, or because you are on
-#   a processor that doesn't have this instruction), ax_cv_gcc_x86_cpuid_OP
-#   is set to the string "unknown".
-#
-#   This macro mainly exists to be used in AX_GCC_ARCHFLAG.
-#
-# LICENSE
-#
-#   Copyright (c) 2008 Steven G. Johnson <stevenj@alum.mit.edu>
-#   Copyright (c) 2008 Matteo Frigo
-#
-#   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, either 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 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/>.
-#
-#   As a special exception, the respective Autoconf Macro's copyright owner
-#   gives unlimited permission to copy, distribute and modify the configure
-#   scripts that are the output of Autoconf when processing the Macro. You
-#   need not follow the terms of the GNU General Public License when using
-#   or distributing such scripts, even though portions of the text of the
-#   Macro appear in them. The GNU General Public License (GPL) does govern
-#   all other use of the material that constitutes the Autoconf Macro.
-#
-#   This special exception to the GPL applies to versions of the Autoconf
-#   Macro released by the Autoconf Archive. When you make and distribute a
-#   modified version of the Autoconf Macro, you may extend this special
-#   exception to the GPL to apply to your modified version as well.
-
-#serial 7
-
-AC_DEFUN([AX_GCC_X86_CPUID],
-[AC_REQUIRE([AC_PROG_CC])
-AC_LANG_PUSH([C])
-AC_CACHE_CHECK(for x86 cpuid $1 output, ax_cv_gcc_x86_cpuid_$1,
- [AC_RUN_IFELSE([AC_LANG_PROGRAM([#include <stdio.h>], [
-     int op = $1, eax, ebx, ecx, edx;
-     FILE *f;
-      __asm__("cpuid"
-        : "=a" (eax), "=b" (ebx), "=c" (ecx), "=d" (edx)
-        : "a" (op));
-     f = fopen("conftest_cpuid", "w"); if (!f) return 1;
-     fprintf(f, "%x:%x:%x:%x\n", eax, ebx, ecx, edx);
-     fclose(f);
-     return 0;
-])],
-     [ax_cv_gcc_x86_cpuid_$1=`cat conftest_cpuid`; rm -f conftest_cpuid],
-     [ax_cv_gcc_x86_cpuid_$1=unknown; rm -f conftest_cpuid],
-     [ax_cv_gcc_x86_cpuid_$1=unknown])])
-AC_LANG_POP([C])
-])

config/ax_sse.m4

@@ -4,14 +4,13 @@
 #
 # SYNOPSIS
 #
-#   AX_EXT
+#   AX_SSE
 #
 # DESCRIPTION
 #
-#   Find supported SIMD extensions by requesting cpuid. When an SIMD
+#   Find SIMD extensions supported by compiler. The -m"simdextensionname" is
-#   extension is found, the -m"simdextensionname" is added to SIMD_FLAGS if
+#   added to SIMD_FLAGS if compiler supports it. For example, if "sse2" is
-#   compiler supports it. For example, if "sse2" is available, then "-msse2"
+#   available, then "-msse2" is added to SIMD_FLAGS.
-#   is added to SIMD_FLAGS.
 #
 #   This macro calls:
 #
@@ -19,7 +18,7 @@
 #
 #   And defines:
 #
-#     HAVE_MMX / HAVE_SSE / HAVE_SSE2 / HAVE_SSE3 / HAVE_SSSE3 / HAVE_SSE4.1 / HAVE_SSE4.2 / HAVE_AVX
+#     HAVE_SSE3 / HAVE_SSE4.1
 #
 # LICENSE
 #
@@ -41,32 +40,29 @@ AC_DEFUN([AX_SSE],
 [
   AC_REQUIRE([AC_CANONICAL_HOST])
 
+  AM_CONDITIONAL(HAVE_SSE3, false)
+  AM_CONDITIONAL(HAVE_SSE4_1, false)
+
   case $host_cpu in
     i[[3456]]86*|x86_64*|amd64*)
-
-      AC_REQUIRE([AX_GCC_X86_CPUID])
-      AC_REQUIRE([AX_GCC_X86_AVX_XGETBV])
-
-      AX_GCC_X86_CPUID(0x00000001)
-
       AX_CHECK_COMPILE_FLAG(-msse3, ax_cv_support_sse3_ext=yes, [])
       if test x"$ax_cv_support_sse3_ext" = x"yes"; then
         SIMD_FLAGS="$SIMD_FLAGS -msse3"
-        AC_DEFINE(HAVE_SSE3,,[Support SSE3 (Streaming SIMD Extensions 3) instructions])
+        AC_DEFINE(HAVE_SSE3,,
+          [Support SSE3 (Streaming SIMD Extensions 3) instructions])
         AM_CONDITIONAL(HAVE_SSE3, true)
       else
-        AC_MSG_WARN([Your compiler does not support sse3 instructions, can you try another compiler?])
+        AC_MSG_WARN([Your compiler does not support SSE3 instructions])
-	AM_CONDITIONAL(HAVE_SSE3, false)
       fi
 
       AX_CHECK_COMPILE_FLAG(-msse4.1, ax_cv_support_sse41_ext=yes, [])
       if test x"$ax_cv_support_sse41_ext" = x"yes"; then
         SIMD_FLAGS="$SIMD_FLAGS -msse4.1"
-        AC_DEFINE(HAVE_SSE4_1,,[Support SSE4.1 (Streaming SIMD Extensions 4.1) instructions])
+        AC_DEFINE(HAVE_SSE4_1,,
+          [Support SSE4.1 (Streaming SIMD Extensions 4.1) instructions])
         AM_CONDITIONAL(HAVE_SSE4_1, true)
       else
-        AC_MSG_WARN([Your compiler does not support sse4.1 instructions, can you try another compiler?])
+        AC_MSG_WARN([Your compiler does not support SSE4.1])
-	AM_CONDITIONAL(HAVE_SSE4_1, false)
       fi
   ;;
   esac

configure.ac

@@ -18,7 +18,9 @@ dnl You should have received a copy of the GNU General Public License
 dnl along with this program.  If not, see <http://www.gnu.org/licenses/>.
 dnl
 
-AC_INIT(openbts,P2.8TRUNK)
+AC_INIT([osmo-trx],
+	m4_esyscmd([./git-version-gen .tarball-veresion]),
+	[openbsc@lists.osmocom.org])
 AC_PREREQ(2.57)
 AC_CONFIG_SRCDIR([Transceiver52M/Makefile.am])
 AC_CONFIG_AUX_DIR([.])	
@@ -34,8 +36,13 @@ AM_INIT_AUTOMAKE([subdir-objects])
 dnl Linux kernel KBuild style compile messages
 m4_ifdef([AM_SILENT_RULES], [AM_SILENT_RULES([yes])])
 
+dnl include release helper
+RELMAKE='-include osmo-release.mk'
+AC_SUBST([RELMAKE])
+
 AM_PROG_AS
 AC_PROG_CXX
+AX_CXX_COMPILE_STDCXX_11
 AC_PROG_LN_S
 AC_PROG_MAKE_SET
 AC_PROG_INSTALL
@@ -120,6 +127,33 @@ AS_IF([test "x$with_sse" != "xno"], [
     AM_CONDITIONAL(HAVE_SSE4_1, false)
 ])
 
+dnl Check if the compiler supports specified GCC's built-in function
+AC_DEFUN([CHECK_BUILTIN_SUPPORT], [
+  AC_CACHE_CHECK(
+    [whether ${CC} has $1 built-in],
+    [osmo_cv_cc_has_builtin], [
+      AC_LINK_IFELSE([
+        AC_LANG_PROGRAM([], [
+          __builtin_cpu_supports("sse");
+        ])
+      ],
+      [AS_VAR_SET([osmo_cv_cc_has_builtin], [yes])],
+      [AS_VAR_SET([osmo_cv_cc_has_builtin], [no])])
+    ]
+  )
+
+  AS_IF([test yes = AS_VAR_GET([osmo_cv_cc_has_builtin])], [
+    AC_DEFINE_UNQUOTED(AS_TR_CPP(HAVE_$1), 1,
+      [Define to 1 if compiler has the '$1' built-in function])
+  ], [
+    AC_MSG_WARN($2)
+  ])
+])
+
+dnl Check if the compiler supports runtime SIMD detection
+CHECK_BUILTIN_SUPPORT([__builtin_cpu_supports],
+  [Runtime SIMD detection will be disabled])
+
 AM_CONDITIONAL(USRP1, [test "x$with_usrp1" = "xyes"])
 AM_CONDITIONAL(ARCH_ARM, [test "x$with_neon" = "xyes" || test "x$with_neon_vfpv4" = "xyes"])
 AM_CONDITIONAL(ARCH_ARM_A15, [test "x$with_neon_vfpv4" = "xyes"])

debian/changelog

@@ -1,3 +1,170 @@
+osmo-trx (0.2.0) unstable; urgency=medium
+
+  [ Alexander Chemeris ]
+  * EDGE: Add support for UmTRX.
+  * Common: Get rid of a compilation warning.
+  * Common: Make sure gLogEarly() log to the same facilities as the normal log.
+  * transceiver: Properly handle MAXDLY.
+  * transceiver: Add an option to generate random Access Bursts.
+  * osmo-trx: Output Rx SPS as a part of configuration output.
+  * transceiver: Do not pass transceiver state struct to function where it's not used.
+  * makefile: Fix build from an external path.
+  * radioDevice: GSMRATE macro must have parentheses around its definition.
+  * uhd: Fix comment.
+  * radioInterface: Initialize power scale with a meaningful default.
+  * transceiver: Log channel number in DEBUG output of demoded bursts.
+  * transceiver: Add an option to emulate a RACH delay in random filler mode.
+  * UHD: Initial LimeSDR support.
+  * CommonLibs: Remove unused files.
+  * sigProcLib: Typo sybols -> symbols
+  * radioBuffer: Remove extra ; at the end of inline function definitions.
+  * sigProcLib: Fix documentation, sync argument names in .cpp and .h files.
+  * sigProcLib: make energyDetect() simpler by returning actual energy.
+  * sigProcLib: Rename demodulateBurst() to demodGmskBurst() for clarity.
+  * sigProcLib: Slice SoftVector instead of signalVector for GMSK demod.
+  * Call vectorSlicer() right before packing bits for transmission to osmo-bts.
+  * CommonLibs: Print soft bits with less confidence to console when printing a soft vector.
+  * BitVector: Remove convolutional codec - we don't use it in osmo-trx.
+  * BitVector: Convert SoftVector from 0..1 to -1..+1 soft bits.
+  * signalVector: Implement segment().
+  * vector: Introduce segmentMove() method to move data inside of a vector.
+  * vector: Introduce shrink() function to shrink vector size without loosing data.
+  * Move CorrType type from Transceiver to sigProcLib.
+  * sigProcLib: rename signalError type to SignalError.
+  * Move Transceiver::detectBurst() to sigProcLib to make it reusable.
+  * Move BURST_THRESH from Transceiver.cpp to sigProcLib.h to make it reusable.
+  * sigProcLib: Add operator<< to print CorrType to a string.
+  * sigProcLib.h: Fix whitespaces. No non-whitespace changes.
+  * Move Transceiver::demodulate() to sigProcLib to make it reusable.
+  * sigProcLib: constify signalVector arguments for detectBurst() functions.
+  * sigProcLib: Constify demodulation functions burst argument.
+  * sigProcLib: Fix number of tail bits in random Normal Bursts and zero Stealing Bits.
+  * Configuration: Variables allocated with 'new' must be freed with 'delete'.
+  * BitVector: Remove Generator class.
+  * PRBS: a Pseudo-random binary sequence (PRBS) generator class.
+
+  [ Tom Tsou ]
+  * EDGE: Fix USRP B210 device support
+  * uhd: Correct timing alignment in 8-PSK and GMSK downlink bursts
+  * EDGE: Fix demodulation slicer input
+  * common: Restrict UDP binding to localhost only
+  * common: Add mandatory length field to UDP receive calls
+  * uhd: Update default E3XX settings
+  * uhd: Set default Tx sampling to 4 sps
+  * uhd: Make device offset check a private method
+  * uhd: Set minimum UHD version requirement for E3XX
+  * sigproc: Expand RACH, TSC, and EDGE correlation windows
+  * transceiver: Do not report error on SETTSC when radio is on
+  * transceiver: Add Rx samples-per-symbol option
+  * radioInterface: Convert diversity argument to general type
+  * iface: Add inner ring-buffer implementation
+  * mcbts: Add multi-ARFCN channelizing filters
+  * mcbts: Add multi-ARFCN radio support
+  * sigproc: Adjust burst detection threshold criteria
+  * egprs: Enable 8-PSK length vectors on the Tx interface
+  * egprs: Enable 8-PSK burst detection when EDGE is enabled
+  * transceiver: Remove HANDOVER warnings
+  * mcbts: Allow out of order channel setup
+  * radioInterface: Fix multi-channel buffer index bug
+  * uhd: Add command line option for GPS reference
+  * transceiver: Fix mixed GSMK / 8-PSK transmission
+  * transceiver: Fix 4 SPS receive TOA value
+  * sigproc: Fix missing 8-PSK tail symbols
+  * uhd: Update USRP2/N200/N210 for 4 SPS Rx
+  * sigproc: Match differential GMSK start/end bits to tail bits
+  * uhd: Add missing B200 sample timing for 4 SPS receive
+  * transceiver: Fix command build warning
+  * uhd: Set minimum supported version to 3.9.0
+  * uhd: Add X300 sample timing for 4 SPS
+  * Revert "uhd: Set minimum supported version to 3.9.0"
+  * uhd: Add support for UHD-3.11 logging control
+  * uhd: Increase MC-BTS FPGA clock rate to 51.2 MHz
+  * Resampler: Fix initialization return checking
+  * sigProcLib: Remove unreachable code and no-effect checks
+  * sigProcLib: Check return status on downsampling
+  * sigProcLib: Fix negative value check on unsigned value
+  * Resampler: Fix non-array delete for filter taps
+  * Transceiver: Remove unsigned negative compares
+  * Configuration: Fix const and signedness compile warnings
+  * config: Remove OpenBTS style sqlite configuration
+  * radioInterface: Remove UmTRX 'diversity' option
+  * build: Require and check for gcc C++11 support
+  * uhd: Use map container for for device parameter access
+  * sigProcLib: Remove unused functions from public interface
+  * uhd: Add non-UmTRX channel swap support
+  * uhd: Fix Tx-RX timing offset setting
+  * uhd: Fix USRP2/N200/N210 device detection
+  * transceiver: Fix POWEROFF crash on USRP2/N200/X300 devices
+  * sigProcLib: Fix complex/real vector flag in Laurent modulator
+  * sigProcLib: Remove heap based signal vector allocations
+  * common: Declare explicit Vector move constructor
+  * sigProcLib: Remove trigonometric tables
+  * sigProcLib: Use explicit NaN check in sinc table generation
+  * sigProcLib: Replace dynamically allocated resampling buffers
+  * sigProcLib: Specify standard namespace for isnan()
+  * uhd: Always specify samples-per-symbol for device lookup
+  * LimeSDR: set approximate tx offset value to make GSM work
+
+  [ Neels Hofmeyr ]
+  * add basic .gitignore
+  * configure.ac: check for boost/config.hpp header
+  * The INSTALL file is being overwritten by autoreconf, but it is committed as empty file. As a result, the INSTALL file always shows as modified. Instead, remove INSTALL from git and ignore it.
+  * add contrib/jenkins.sh, for gerrit build bot
+
+  [ pierre.baudry ]
+  * transceiver: Fix mismatched allocations and deallocations
+
+  [ Holger Hans Peter Freyther ]
+  * debian: Require fftw3 header files for osmo-trx
+
+  [ Max ]
+  * Add gerrit settings
+  * Integrate Debian packaging changes
+  * Remove embedded sqlite3
+  * Fix building against sqlite3
+  * Add autoconf-archive to dependencies
+  * debian: remove obsolete dependency
+  * deb: remove unused dependency
+  * Remove redundant explicit dependency
+  * Use release helper from libosmocore
+
+  [ Ruben Undheim ]
+  * Do not embed sqlite3 when building
+
+  [ Philipp Maier ]
+  * buildenv: Turn off native architecture builds
+  * cosmetic: Make parameter lists uniform
+  * Add test program to verify convolution implementation
+  * ssedetect: Add runtime CPU detection
+  * cosmetic: remove code duplication
+  * buildenv: Make build CPU invariant
+  * buildenv: Split up SSE3 and SSE4.1 code
+  * cosmetic: Add info about SSE support
+
+  [ Vadim Yanitskiy ]
+  * buildenv: correct the ax_sse macro description
+  * buildenv: actually strip unused cpuid functionality
+  * buildenv: fix build on systems without SIMD support
+  * buildenv: cosmetic changes
+  * buildenv: check for __builtin_cpu_supports call support
+  * ssedetect: call __builtin_cpu_supports() only if supported
+
+  [ Pau Espin Pedrol ]
+  * cosmetic: transciever: Remove trailing whitespaces
+  * transceiver: Avoid sending clock indications when trx is not powered on
+  * Add -j option to bind to specific address
+
+  [ ignasj ]
+  * LimeSDR: Change device detection to work with USB and PCIe versions
+  * LimeSDR: change tx window type to TX_WINDOW_FIXED
+  * LimeSDR: Fix sample value range
+
+  [ Harald Welte ]
+  * Add '-t' command line option to enable SCHED_RR
+  * Import git-version-gen and update AC_INIT()
+
+ -- Harald Welte <laforge@gnumonks.org>  Sat, 28 Oct 2017 17:52:32 +0200
+
 osmo-trx (0.1.9) trusty; urgency=medium
 
   * Ask Ivan, really

debian/control

@@ -5,7 +5,6 @@ Maintainer: Ivan Klyuchnikov <ivan.kluchnikov@fairwaves.ru>
 Build-Depends: debhelper (>= 9),
                autotools-dev,
                autoconf-archive,
-               libdbd-sqlite3,
                libsqlite3-dev,
                pkg-config,
                dh-autoreconf,
@@ -20,7 +19,7 @@ Homepage: https://projects.osmocom.org/projects/osmotrx
 
 Package: osmo-trx
 Architecture: any
-Depends: ${shlibs:Depends}, ${misc:Depends}, libdbd-sqlite3
+Depends: ${shlibs:Depends}, ${misc:Depends}
 Description: SDR transceiver that implements Layer 1 of a GSM BTS
  OsmoTRX is a software-defined radio transceiver that implements the Layer 1
  physical layer of a BTS comprising the following 3GPP specifications:

git-version-gen

@@ -0,0 +1,151 @@
+#!/bin/sh
+# Print a version string.
+scriptversion=2010-01-28.01
+
+# Copyright (C) 2007-2010 Free Software Foundation, Inc.
+#
+# 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; either 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 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/>.
+
+# This script is derived from GIT-VERSION-GEN from GIT: http://git.or.cz/.
+# It may be run two ways:
+# - from a git repository in which the "git describe" command below
+#   produces useful output (thus requiring at least one signed tag)
+# - from a non-git-repo directory containing a .tarball-version file, which
+#   presumes this script is invoked like "./git-version-gen .tarball-version".
+
+# In order to use intra-version strings in your project, you will need two
+# separate generated version string files:
+#
+# .tarball-version - present only in a distribution tarball, and not in
+#   a checked-out repository.  Created with contents that were learned at
+#   the last time autoconf was run, and used by git-version-gen.  Must not
+#   be present in either $(srcdir) or $(builddir) for git-version-gen to
+#   give accurate answers during normal development with a checked out tree,
+#   but must be present in a tarball when there is no version control system.
+#   Therefore, it cannot be used in any dependencies.  GNUmakefile has
+#   hooks to force a reconfigure at distribution time to get the value
+#   correct, without penalizing normal development with extra reconfigures.
+#
+# .version - present in a checked-out repository and in a distribution
+#   tarball.  Usable in dependencies, particularly for files that don't
+#   want to depend on config.h but do want to track version changes.
+#   Delete this file prior to any autoconf run where you want to rebuild
+#   files to pick up a version string change; and leave it stale to
+#   minimize rebuild time after unrelated changes to configure sources.
+#
+# It is probably wise to add these two files to .gitignore, so that you
+# don't accidentally commit either generated file.
+#
+# Use the following line in your configure.ac, so that $(VERSION) will
+# automatically be up-to-date each time configure is run (and note that
+# since configure.ac no longer includes a version string, Makefile rules
+# should not depend on configure.ac for version updates).
+#
+# AC_INIT([GNU project],
+#         m4_esyscmd([build-aux/git-version-gen .tarball-version]),
+#         [bug-project@example])
+#
+# Then use the following lines in your Makefile.am, so that .version
+# will be present for dependencies, and so that .tarball-version will
+# exist in distribution tarballs.
+#
+# BUILT_SOURCES = $(top_srcdir)/.version
+# $(top_srcdir)/.version:
+#	echo $(VERSION) > $@-t && mv $@-t $@
+# dist-hook:
+#	echo $(VERSION) > $(distdir)/.tarball-version
+
+case $# in
+    1) ;;
+    *) echo 1>&2 "Usage: $0 \$srcdir/.tarball-version"; exit 1;;
+esac
+
+tarball_version_file=$1
+nl='
+'
+
+# First see if there is a tarball-only version file.
+# then try "git describe", then default.
+if test -f $tarball_version_file
+then
+    v=`cat $tarball_version_file` || exit 1
+    case $v in
+	*$nl*) v= ;; # reject multi-line output
+	[0-9]*) ;;
+	*) v= ;;
+    esac
+    test -z "$v" \
+	&& echo "$0: WARNING: $tarball_version_file seems to be damaged" 1>&2
+fi
+
+if test -n "$v"
+then
+    : # use $v
+elif
+       v=`git describe --abbrev=4 --match='v*' HEAD 2>/dev/null \
+	  || git describe --abbrev=4 HEAD 2>/dev/null` \
+    && case $v in
+	 [0-9]*) ;;
+	 v[0-9]*) ;;
+	 *) (exit 1) ;;
+       esac
+then
+    # Is this a new git that lists number of commits since the last
+    # tag or the previous older version that did not?
+    #   Newer: v6.10-77-g0f8faeb
+    #   Older: v6.10-g0f8faeb
+    case $v in
+	*-*-*) : git describe is okay three part flavor ;;
+	*-*)
+	    : git describe is older two part flavor
+	    # Recreate the number of commits and rewrite such that the
+	    # result is the same as if we were using the newer version
+	    # of git describe.
+	    vtag=`echo "$v" | sed 's/-.*//'`
+	    numcommits=`git rev-list "$vtag"..HEAD | wc -l`
+	    v=`echo "$v" | sed "s/\(.*\)-\(.*\)/\1-$numcommits-\2/"`;
+	    ;;
+    esac
+
+    # Change the first '-' to a '.', so version-comparing tools work properly.
+    # Remove the "g" in git describe's output string, to save a byte.
+    v=`echo "$v" | sed 's/-/./;s/\(.*\)-g/\1-/'`;
+else
+    v=UNKNOWN
+fi
+
+v=`echo "$v" |sed 's/^v//'`
+
+# Don't declare a version "dirty" merely because a time stamp has changed.
+git status > /dev/null 2>&1
+
+dirty=`sh -c 'git diff-index --name-only HEAD' 2>/dev/null` || dirty=
+case "$dirty" in
+    '') ;;
+    *) # Append the suffix only if there isn't one already.
+	case $v in
+	  *-dirty) ;;
+	  *) v="$v-dirty" ;;
+	esac ;;
+esac
+
+# Omit the trailing newline, so that m4_esyscmd can use the result directly.
+echo "$v" | tr -d '\012'
+
+# Local variables:
+# eval: (add-hook 'write-file-hooks 'time-stamp)
+# time-stamp-start: "scriptversion="
+# time-stamp-format: "%:y-%02m-%02d.%02H"
+# time-stamp-end: "$"
+# End: