Model for an F Engine used by the Precision Array for Probing the Epoch of
Reionization.

The Model roachf_1024ch_ma.mdl is an 8 input design consisting of Digitizer,
PFB, Packetiser, Loopback, and output 10gbe.  2048 channel PFB. 

One can select the total number of inputs in the design (which determines the
packet format).  Supported number of inputs is from 16, 32, 64, 128, 256, 512
total inputs. 

This mdl file was built using Xilinix 13.3 toolflow and 
casper mlib devel library at #4655b346c5f629aa63ed5e50d38a75be4ef33102.
mlib devel can be found at [email protected]:zakiali/mlib_devel.git

REGISTERS OF NOTE
------------------

	Monitor
	-------
	- 'switch_gbe_status' : monitors 10gbe outputs. 
	   	bitmap = {25bits of 0 | 
			  gbe_up | 
			  gbe_rx | 
			  gbe_tx | 
			  gbe_a_full |
			  gbe_tx_oflow | 
			  gbe_rx_badframe | 
 			  gbe_rx_over_run}
	
	- 'switch_gbe_badframe' : Counter value of gbe rx bad frame signal.
				  Note there is no posedge.  

	- 'switch_gbe_orun' : Counter value of gbe rx overrun signal. Note
		              there is no posedge.  

	- 'switch_gbe_oflow' : Counter value of gbe tx over flow signal. Note
			       there is no posedge.  

	- 'switch_cnts_data_msb(lsb)' : The cnts block counts the number of
					clocks between packets etc.. use
					provided software to read.  May be
					somewhat convoluted to decipher. Not
					needed for functionality. Only for
					testing 10gbe blocks. 

	- 'loopback_rx_cnt' : Counts the number of packets output of rx packet
		              fifo. Reset by 'cnt_rst' signal.

	- 'loopback_rx_err_cnt' : Counts the number of rx fifo overfull
				  signals. Reset by 'cnt_rst' signal.

	- 'loopback_loop_err_cnt' : Counts the number of loopback fifo overfull
				    signals. Reset by 'cnt_rst' signal.

	- 'loopback_loop_cnt' : Counts the number of packets output of loop
				packet fifo. Reset by 'cnt_rst' signal.  This
				number should be 
				(1/number of roaches) * loopback_rx_cnt.
	
	- 'loopback_rx_pkt_too_big' : Count of number to oversize packets
			              coming from the switch. 

	- 'loopback_rx_pkt_too_small' : Count of number to undersize packets
				        coming from the switch. 

	- 'loopback_rx_over' : Second check of oversize packets from switch.
				  NOT NECESSARY.

	- 'loopback_rx_under' : Second check of undersize packets from switch.
				  NOT NECESSARY.

	- 'loopback_loop_over' : Count of number of oversize packets from
				 loopback fifo.

	- 'loopback_loop_under' : Count of number of undersize packets from
				  loopback fifo.

	- 'loopback_cnts.,,,' : Same as cnts brams in switch block. Counts
				differences between gbe_read and loop_read
				signals.  Was used for debugging. Not needed
				anymore.

	- 'gpu_mcnt_lsb(msb)' : Instantaneous header  of data about to be sent
				out over 10gbe. LSB = {bottom 16 of mcnt |
				Xid | Fid} MSB = top 32 bits of mcnt.

	- 'gpu_gbe_status' : see 'switch_gbe_status.'

	- 'adc_rms_levels' : bram of rms values of the inputs. Note that the
			     ctrl register must be written to in order for it
			     to run. addresses: 
				0: 24bit sum of 2^16 samples of input 0. 
				1: 32bit sum of squares of the samples in 0.
				2: 24bit sum of 2^16 samples of input 1.
				3: 32bit sum of squares of the samples in 2.
				etc.....





	LEDs
	------
	
	led0 : armed signal

	led1 : sync pulse 

	led2 : eq clip

	led3 : adc clip








	SNAP BLOCKS
	-------------
	
	-'switch_txsnap' + 'bram_msb', 'bram_lsb' , 'bram_oob', 'ctrl', 'addr':
		saves meta data in oob, lsb(msb) of data. Triggered by writing
		to ctrl signal.  See scripts to read from it.

	-'switch_rxsnap' + 'bram_msb', 'bram_lsb' , 'bram_oob', 'ctrl', 'addr':
		saves meta data in oob, lsb(msb) of data. Triggered by writing
		to ctrl signal.  See scripts to read from it.

	-'gpu_txsnap' + 'bram_msb', 'bram_lsb' , 'bram_oob', 'ctrl', 'addr':
		saves meta data in oob, lsb(msb) of data. Triggered by writing
		to ctrl signal.  See scripts to read from it.

	-'adc_snap_0_3'(4_7) : use 'adc_snap_ctrl' to control snaps.
		1024 samples of 8bit input data. 
		Format ={[0123]_0,[0123]_0...,[0123]_1023} 
			and {[4567]_0,[4567]_1,...,[4567]_1023}









	CONTROL BLOCKS
	---------------

	- 'input_selector' : Selects the type of input. 32bit value. Every
	  nibble corresponds to an input. lsb is input 0, msb is input 7. If
	  writtin in hex format, every hex digit is an input. value of
	  3=digital zero, 2=second digital noise, 1=first digital noise, 0=adc
	  inputs.  e.g. 0x33333210 = input0 has adcs on, input 1 has digital
	  noise 1, input 2 has digital noise 2 on , inputs3-7 have dig zero.
	
	- 'delay_values' : introduces a course delay into input signal path.
	  can delay up to 16 fpga clocks. lsb is input 0, msb is input 7.

	- 'seed_data' : introduces new seed into digital noise sources. every
	  byte corresponds to a different wgn generator. note that this value
	  must be set before we arm/sync up the roaches to one another.

	- 'gpu_port' : destination port for  gpu host machine. default is 8511.

	- 'gpu_ip' : destination ip address for gpu host machine.

	- 'gbe_sw_port' : destination port for packets destined for switch.

	- 'my_ip' : ipaddress of this f engine.

	- 'ip_base' : base ip address of packets destined for switch.
				
	- 'ant_base_offset' : This is actually the F-Engine ID. Determines
	  which input numbers are on this roach.
	
	- 'fft_shift' : sets shift schedule for ffts.

	- 'n_inputs' : Register holds the number of total inputs in the
	  correlator. This determines which transpose/packetiser to use. It is
	  a 3 bit value.  The number is determined by log2(ninputs) - 4.

	- 'feng_ctl'* : sets various control bits. Bit map below...
		('gbe_gpu_rst'),        #31                reset gbe block going
							   to gpu machines.

		('gbe_sw_rst'),         #30                reset gbe block going
							   to gpu machines.

		('loopback_mux_rst'),   #29                reset some loopback
							   specific stuff.
							   Currently not used,
							   except for
							   determining proper
							   header in mux.

		('cnt_rst'),            #28                resets various
							   counters (error and
							   the such) in the
							   design.

		('fft_preshift',2), 	#26-27             can preshift (down)
							   into fft by up to 3
							   bits. 0=noshift(def)
							   1=1 bit, etc... 

		('gpio_monsel',3),  	#23,24,25          DEPRECATED. Not used 
						           at all.

		('fft_tvg2'),           #22                use fft test vector 
							   generator 2. Don't
							   really use this.
							   Digital noise
							   sources are better.

		('fft_tvg1'),           #21		   use fft test vector 
							   generator 2. Don't
							   really use this.
							   Digital noise
							   sources are better.

		('gbe_gpu_disable'),    #20                Disable gpu 10gbe 
							   block. (Valid signal
							   always low)

		('use_qdr_tvg'),        #19                (actually called
							   use_sram_tvg). Use
							   post transpose test
							   vector. Never Used.

		('gbe_sw_disable'),     #18                Disable switch 10gbe
							   block. (Valid signal

							   always low)
		('arm_rst'),            #17                Rearm the roachs
							   (for syncup). Note
							   this must go from 0
							   to 1, for rearm to
							   take effect.
							   
		('sync_rst'),           #16                Reset sync pulse 
							   manually. This is
							   never used, since we
							   have a 1pps
							   distributor.

		Padding(13),            #3-15              nada. 

		('lb_err_cnt_rst'),     #2                 Reset the cnts
							   blocks in switch and
							   loopback.

		Padding(2))             #0-1               nada.

* Some notes about this block :  The 'gbe_gpu_disable' and
				 'gbe_sw_disable' signals need to be high
				 before the 10gbe cores are reset with. If not,
				 the cores will not reset and will not
				 transmit.  Also the cores are held in rst
				 until all the network stuff is set up. Then
				 the final start up of the F Engine is to take
				 the 10gbe cores out of reset and and undo the
				 disable. The gbe disable signals make sure
				 that no valid data goes into the cores, and
				 therefore no data gets sent to the 10gbe
				 cores.


	NOTES
       =======
	
	1. The 2 10gbe cores in the design are different. The core to the
	   switch is the 10gbe_v2 core, where as the core to the GPU's is the
	   10gbe (original) core. It requires a special liscense from Xilinx to
	   be compiled with.

	2. In the initialization scripts, which can be found in the corr
	   package at [email protected]:zakiali/PAPERCORR.git, both the 10gbe
	   cores are started the same way (as in I don't use tgtap to start
	   them up). This was due to the fact that sometimes we did not have
	   all ROACHs set up for testing, but we were sending all the packets
	   to the switch. The packets that were destined for non existant
	   ROACHs were then interpreted as broadcast packets and sent
	   everywhere, corrupting data integrity.  Therefore, I populate the
	   arp tables manually, including all the mac addresses that are
	   supposed to be there (non existant ROACHs).

	3. The name of the gbe cores relates to the cx4 ports used on the 
	   roachs. The number in the name is the port used (starting from 0).