Changed gr-gsm::receiver block into typical typical sink inherited from sync_block.
Removed forecast(...) function and replaced it with call to set_output_multiple in the constructor.
diff --git a/lib/receiver_impl.cc b/lib/receiver_impl.cc
index 71a9494..6e92c2d 100644
--- a/lib/receiver_impl.cc
+++ b/lib/receiver_impl.cc
@@ -62,7 +62,7 @@
* The private constructor
*/
receiver_impl::receiver_impl(feval_dd * tuner, int osr, int arfcn)
- : gr::block("receiver",
+ : gr::sync_block("receiver",
gr::io_signature::make(1, 1, sizeof(gr_complex)),
gr::io_signature::make(0, 0, 0)),
d_OSR(osr),
@@ -78,6 +78,8 @@
d_signal_dbm(-120)
{
int i;
+ set_output_multiple(floor((TS_BITS + 2 * GUARD_PERIOD) * d_OSR)); //don't send samples to the receiver until there are at least samples for one
+ // burst and two gurad periods (one gurard period is an arbitrary overlap)
gmsk_mapper(SYNC_BITS, N_SYNC_BITS, d_sch_training_seq, gr_complex(0.0, -1.0));
for (i = 0; i < TRAIN_SEQ_NUM; i++)
{
@@ -95,18 +97,12 @@
{
}
-void receiver_impl::forecast(int noutput_items, gr_vector_int &ninput_items_required)
-{
- ninput_items_required[0] = noutput_items * floor((TS_BITS + 2 * GUARD_PERIOD) * d_OSR);
-}
-
-
int
-receiver_impl::general_work(int noutput_items,
- gr_vector_int &ninput_items,
- gr_vector_const_void_star &input_items,
- gr_vector_void_star &output_items)
+receiver_impl::work(int noutput_items,
+ gr_vector_const_void_star &input_items,
+ gr_vector_void_star &output_items)
{
+ //std::cout << noutput_items << std::endl;
const gr_complex *input = (const gr_complex *) input_items[0];
switch (d_state)
@@ -114,7 +110,7 @@
//bootstrapping
case first_fcch_search:
DCOUT("FCCH search");
- if (find_fcch_burst(input, ninput_items[0])) //find frequency correction burst in the input buffer
+ if (find_fcch_burst(input, noutput_items)) //find frequency correction burst in the input buffer
{
//set_frequency(d_freq_offset); //if fcch search is successful set frequency offset
COUT("Freq offset " << d_freq_offset);
@@ -130,7 +126,7 @@
{
DCOUT("NEXT FCCH search");
float prev_freq_offset = d_freq_offset; //before previous set_frequqency cause change
- if (find_fcch_burst(input, ninput_items[0]))
+ if (find_fcch_burst(input, noutput_items))
{
if (abs(prev_freq_offset - d_freq_offset) > FCCH_MAX_FREQ_OFFSET)
{
@@ -155,7 +151,7 @@
int burst_start = 0;
unsigned char output_binary[BURST_SIZE];
- if (reach_sch_burst(ninput_items[0])) //wait for a SCH burst
+ if (reach_sch_burst(noutput_items)) //wait for a SCH burst
{
burst_start = get_sch_chan_imp_resp(input, &channel_imp_resp[0]); //get channel impulse response from it
detect_burst(input, &channel_imp_resp[0], burst_start, output_binary); //detect bits using MLSE detection
@@ -192,7 +188,7 @@
burst_type b_type = d_channel_conf.get_burst_type(d_burst_nr); //get burst type for given burst number
double signal_pwr=0;
- for(int ii=0;ii<ninput_items[0];ii++)
+ for(int ii=0;ii<noutput_items;ii++)
{
signal_pwr += abs(input[ii])*abs(input[ii]);
}
@@ -769,14 +765,11 @@
}
*corr_max = max_correlation;
- // We want to use the first sample of the impulse response, and the
- // corresponding samples of the received signal.
- // the variable sync_w should contain the beginning of the used part of
- // training sequence, which is 3+57+1+6=67 bits into the burst. That is
- // we have that sync_t16 equals first sample in bit number 67.
DCOUT("strongest_window_nr_new: " << strongest_window_nr);
- burst_start = search_start_pos + strongest_window_nr - TRAIN_POS * d_OSR;
+ burst_start = search_start_pos + strongest_window_nr - TRAIN_POS * d_OSR; //compute first sample posiiton which corresponds to the first sample of the impulse response
+ //TRAIN_POS=3+57+1+6
+ //TODO: describe this part in detail in documentation as this is crucial part for synchronization
DCOUT("burst_start: " << burst_start);
return burst_start;