Merge pull request drowe67#242 from drowe67/ms-arm-flex6k-nomerge

Embedded ARM fixes and optimizations

Author: drowe67

CMakeLists.txt

@@ -112,7 +112,7 @@ else()
 endif()
 
 # -fPIC is implied on MinGW...
-if(NOT WIN32)
+if((NOT WIN32) AND (NOT MICROCONTROLLER_BUILD))
     set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -fPIC")
 endif()
 

README.md

@@ -236,6 +236,43 @@ Codec 2 can be added to the project in the following way.
      
 1. Add Codec 2 to the target_link_libraries in the same file.
 
+## Building Codec 2 for Microcontrollers
+
+Codec 2 requires a hardware Floating Point Unit (FPU) to run in real time.
+
+Two build options have been added to support building on microcontrollers:
+1. Setting the `cmake` variable MICROCONTROLLER_BUILD disables position independent code (-fPIC is not used).  This was required for the IMRT1052 used in Teensy 4/4.1).
+
+1. On ARM machines, setting the C Flag \_\_EMBEDDED\_\_ and linking with the ARM CMSIS library will improve performance on ARM-based microcontrollers. \_\_REAL\_\_ and FDV\_ARM\_MATH are additional ARM-specific options that can be set to improve performance if required, especially with OFDM modes.
+
+A CMakeLists.txt example for a microcontroller is below:
+
+```
+set(CMAKE_TRY_COMPILE_TARGET_TYPE STATIC_LIBRARY)
+set(MICROCONTROLLER_BUILD 1)
+
+set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -mlittle-endian -ffunction-sections -fdata-sections -g -O3")
+set(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} -ffunction-sections -fdata-sections")
+
+add_definitions(-DCORTEX_M7 -D__EMBEDDED__)
+add_definitions(-DFREEDV_MODE_EN_DEFAULT=0 -DFREEDV_MODE_1600_EN=1 -DFREEDV_MODE_700D_EN=1 -DFREEDV_MODE_700E_EN=1 -DCODEC2_MODE_EN_DEFAULT=0 -DCODEC2_MODE_1300_EN=1 -DCODEC2_MODE_700C_EN=1)
+                    
+FetchContent_Declare(codec2
+    GIT_REPOSITORY https://github.com/drowe67/codec2.git
+    GIT_TAG origin/master
+    GIT_SHALLOW ON
+    GIT_PROGRESS ON
+)
+FetchContent_GetProperties(codec2)
+if(NOT ${codec2_POPULATED})
+    FetchContent_Populate(codec2)
+endif()
+set(CMAKE_REQUIRED_FLAGS "")
+
+set(LPCNET OFF CACHE BOOL "")
+add_subdirectory(${codec2_SOURCE_DIR} ${codec2_BINARY_DIR} EXCLUDE_FROM_ALL)
+```
+
 ## Building Debian packages
 
 To build Debian packages, simply run the "cpack" command after running "make". This will generate the following packages:

src/codec2.c

@@ -1751,7 +1751,7 @@ float codec2_get_energy(struct CODEC2 *c2, const unsigned char *bits)
     MODEL model;
     float xq_dec[2] = {};
     int e_index, WoE_index;
-    float e;
+    float e = 0.0f;
     unsigned int nbit;
 
     if ( CODEC2_MODE_ACTIVE(CODEC2_MODE_3200, c2->mode)) {

src/codec2_fft.h

@@ -14,10 +14,11 @@
 #include <string.h>
 #include <math.h>
 
-#ifdef FDV_ARM_MATH
-  #include "fdv_arm_math.h"
+#ifndef FDV_ARM_MATH
+  #define USE_KISS_FFT
 #else
-    #define USE_KISS_FFT
+  #include "arm_math.h"
+  #include "arm_const_structs.h"
 #endif
 
 #include "defines.h"

src/fmfsk.c

@@ -116,9 +116,11 @@ void fmfsk_get_demod_stats(struct FMFSK *fmfsk,struct MODEM_STATS *stats){
     stats->rx_timing = fmfsk->stats->rx_timing;
     stats->foff = fmfsk->stats->foff;
 
+#ifndef __EMBEDDED__
     stats->neyesamp = fmfsk->stats->neyesamp;
     stats->neyetr = fmfsk->stats->neyetr;
     memcpy(stats->rx_eye, fmfsk->stats->rx_eye, sizeof(stats->rx_eye));
+#endif // !__EMBEDDED__
 
     /* these fields not used for FSK so set to something sensible */
 
@@ -340,6 +342,7 @@ void fmfsk_demod(struct FMFSK *fmfsk, uint8_t rx_bits[],float fmfsk_in[]){
         fmfsk->snr_mean = 0.9 * fmfsk->snr_mean + 0.1 * (10.0 * log10f(var_signal / var_noise));
     fmfsk->stats->snr_est = fmfsk->snr_mean;
 
+#ifndef __EMBEDDED__
     /* Collect an eye diagram */
     /* Take a sample for the eye diagrams */
     neyesamp = fmfsk->stats->neyesamp = Ts*4;
@@ -361,7 +364,8 @@ void fmfsk_demod(struct FMFSK *fmfsk, uint8_t rx_bits[],float fmfsk_in[]){
     for(i=0; i<fmfsk->stats->neyetr; i++)
         for(j=0; j<neyesamp; j++)
             fmfsk->stats->rx_eye[i][j] = (fmfsk->stats->rx_eye[i][j]/(2*eye_max))+.5;
-    
+#endif // !__EMBEDDED__
+ 
     modem_probe_samp_f("t_norm_rx_timing",&norm_rx_timing,1);
     modem_probe_samp_f("t_rx_filt",rx_filt,(nsym+1)*Ts);
 

src/fsk.c

@@ -867,7 +867,7 @@ void fsk_demod_core(struct FSK *fsk, uint8_t rx_bits[], float rx_filt[], COMP fs
     /* due to oversample rate P, we have too many samples for eye
        trace.  So lets output a decimated version.  We use 2P
        as we want two symbols worth of samples in trace  */
-
+#ifndef __EMBEDDED__
     int neyesamp_dec = ceil(((float)P*2)/MODEM_STATS_EYE_IND_MAX);
     neyesamp = (P*2)/neyesamp_dec;
     assert(neyesamp <= MODEM_STATS_EYE_IND_MAX);
@@ -915,7 +915,8 @@ void fsk_demod_core(struct FSK *fsk, uint8_t rx_bits[], float rx_filt[], COMP fs
 
     for(i=0; i<M; i++)
         fsk->stats->f_est[i] = f_est[i];
-    
+#endif // !__EMBEDDED__
+ 
     /* Dump some internal samples */
     modem_probe_samp_f("t_EbNodB",&(fsk->EbNodB),1);
     modem_probe_samp_f("t_ppm",&(fsk->ppm),1);
@@ -956,7 +957,7 @@ static void stats_init(struct FSK *fsk) {
     /* asserts below as we found some problems over-running eye matrix */
 
     /* TODO: refactor eye tracing code here and in fsk_demod */
-    
+#ifndef __EMBEDDED__    
     int neyesamp_dec = ceil(((float)P*2)/MODEM_STATS_EYE_IND_MAX);
     int neyesamp = (P*2)/neyesamp_dec;
     assert(neyesamp <= MODEM_STATS_EYE_IND_MAX);
@@ -973,6 +974,7 @@ static void stats_init(struct FSK *fsk) {
            }
         }
     }
+#endif // !__EMBEDDED__
 
     fsk->stats->rx_timing = fsk->stats->snr_est = 0;
 
@@ -1006,12 +1008,13 @@ void fsk_get_demod_stats(struct FSK *fsk, struct MODEM_STATS *stats){
     stats->snr_est = fsk->stats->snr_est;           // TODO: make this SNR not Eb/No
     stats->rx_timing = fsk->stats->rx_timing;
     stats->foff = fsk->stats->foff;
-
+#ifndef __EMBEDDED__
     stats->neyesamp = fsk->stats->neyesamp;
     stats->neyetr = fsk->stats->neyetr;
     memcpy(stats->rx_eye, fsk->stats->rx_eye, sizeof(stats->rx_eye));
     memcpy(stats->f_est, fsk->stats->f_est, fsk->mode*sizeof(float));
-        
+#endif // !__EMBEDDED__
+ 
     /* these fields not used for FSK so set to something sensible */
 
     stats->sync = 0;

src/ofdm.c

@@ -45,7 +45,7 @@
 
 #ifdef __EMBEDDED__
 #include "arm_math.h"
-#endif /* __EMBEDDED_ */
+#endif /* __EMBEDDED__ */
 
 /* Static Prototypes */
 
@@ -758,7 +758,7 @@ static int est_timing(struct OFDM *ofdm, complex float *rx, int length,
 	    corr_st = corr_st + (rx[ind                        ] * wvec_pilot[j]);
             corr_en = corr_en + (rx[ind + ofdm->samplesperframe] * wvec_pilot[j]);
         }
-#endif
+#endif // __EMBEDDED__
         corr[i] = (cabsf(corr_st) + cabsf(corr_en)) * av_level;
     }