Skip to content

X
x393
Commit 6989d804 authored by Andrey Filippov's avatar Andrey Filippov
Browse files
Options

working on DQS+DQ inpute delays adjustment

parent 4f1a1c28
Expand all Show whitespace changes
Inline Side-by-side
Showing with 434 additions and 16 deletions
py393/x393_mcntrl_adjust.py
View file @ 6989d804
This diff is collapsed.
py393/x393_mcntrl_timing.py
View file @ 6989d804
......@@ -178,12 +178,14 @@ class X393McntrlTiming(object):
self.axi_set_phase()
def axi_set_dq_idelay(self, # sets same delay to all dq idelay
delay=None): # input [7:0] delay;
delay=None, # input [7:0] delay;
quiet=1):
"""
Set all DQ input delays to the same value
<delay> 8-bit (5+3) delay value to use or a tuple/list with a pair for (lane0, lane1)
Each of the two elements in the delay tuple/list may be a a common integer or a list/tuple itself
if delay is None will restore default values
<quiet> reduce output
"""
# print("====axi_set_dq_idelay %s"%str(delay))
......@@ -194,7 +196,7 @@ class X393McntrlTiming(object):
delay[1].append(vrlg.get_default_field("DLY_LANE1_IDELAY",i))
if isinstance(delay,(int,long)):
delay=(delay,delay)
if self.DEBUG_MODE > 1:
if quiet < 2:
print("SET DQ IDELAY="+hexMultiple(delay)) # hexMultiple
self.axi_set_multiple_delays(vrlg.LD_DLY_LANE0_IDELAY, 0, 8, delay[0], "DLY_LANE0_IDELAY")
self.axi_set_multiple_delays(vrlg.LD_DLY_LANE1_IDELAY, 0, 8, delay[1], "DLY_LANE1_IDELAY")
......
py393/x393_pio_sequences.py
View file @ 6989d804
......@@ -611,11 +611,13 @@ class X393PIOSequences(object):
# Set MR3, read nrep*8 words, save to buffer (port0). No ACTIVATE/PRECHARGE are needed/allowed
def set_read_pattern (self,
nrep, # input integer nrep;
sel=1, # 0 - early, 1 - late read command (shift by a SDCLK period)
verbose=0):
"""
Setup read pattern sequence at parameter defined address in the sequencer memory
<nrep> number of times pattern burst is read
<sel> 0 - early, 1 - late read command (shift by a SDCLK period)
<verbose> print data being written (default: False)
"""
......@@ -639,43 +641,43 @@ class X393PIOSequences(object):
# first read
#@ read
# addr bank RCW ODT CKE SEL DQEN DQSEN DQSTGL DCI B_WR B_RD NOP, B_RST
data=self.func_encode_cmd( 0, 0, 2, 0, 0, 1, 0, 0, 0, 1, 1, 0, 0, 0)
data=self.func_encode_cmd( 0, 0, 2, 0, 0, sel, 0, 0, 0, 1, 1, 0, 0, 0)
self.x393_mem.axi_write_single_w(cmd_addr, data, verbose)
cmd_addr += 1
# nop (combine with previous?)
# skip done bank ODT CKE SEL DQEN DQSEN DQSTGL DCI B_WR B_RD B_RST
data=self.func_encode_skip( 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 1, 0, 0)
data=self.func_encode_skip( 0, 0, 0, 0, 0, sel, 0, 0, 0, 1, 1, 0, 0)
self.x393_mem.axi_write_single_w(cmd_addr, data, verbose)
cmd_addr += 1
#repeat remaining reads
# for (i = 1; i < nrep; i = i + 1) begin
for _ in range(1,nrep):
# addr bank RCW ODT CKE SEL DQEN DQSEN DQSTGL DCI B_WR B_RD NOP, B_RST
data=self.func_encode_cmd( 0, 0, 2, 0, 0, 1, 0, 0, 0, 1, 1, 0, 1, 0)
data=self.func_encode_cmd( 0, 0, 2, 0, 0, sel, 0, 0, 0, 1, 1, 0, 1, 0)
self.x393_mem.axi_write_single_w(cmd_addr, data, verbose)
cmd_addr += 1
# nop - all 3 below are the same? - just repeat?
# skip done bank ODT CKE SEL DQEN DQSEN DQSTGL DCI B_WR B_RD B_RST
data=self.func_encode_skip( 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0)
data=self.func_encode_skip( 0, 0, 0, 0, 0, sel, 0, 0, 0, 1, 0, 0, 0)
self.x393_mem.axi_write_single_w(cmd_addr, data, verbose)
cmd_addr += 1
# nop
# skip done bank ODT CKE SEL DQEN DQSEN DQSTGL DCI B_WR B_RD B_RST
data=self.func_encode_skip( 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0)
data=self.func_encode_skip( 0, 0, 0, 0, 0, sel, 0, 0, 0, 1, 0, 0, 0)
self.x393_mem.axi_write_single_w(cmd_addr, data, verbose)
cmd_addr += 1
# nop
# skip done bank ODT CKE SEL DQEN DQSEN DQSTGL DCI B_WR B_RD B_RST
data=self.func_encode_skip( 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0)
data=self.func_encode_skip( 0, 0, 0, 0, 0, sel, 0, 0, 0, 1, 0, 0, 0)
self.x393_mem.axi_write_single_w(cmd_addr, data, verbose)
cmd_addr += 1
# nop, no write buffer - next page
# skip done bank ODT CKE SEL DQEN DQSEN DQSTGL DCI B_WR B_RD B_RST
data=self.func_encode_skip( 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 1)
data=self.func_encode_skip( 0, 0, 0, 0, 0, sel, 0, 0, 0, 1, 0, 0, 1)
self.x393_mem.axi_write_single_w(cmd_addr, data, verbose)
cmd_addr += 1
# skip done bank ODT CKE SEL DQEN DQSEN DQSTGL DCI B_WR B_RD B_RST
data=self.func_encode_skip( 1, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0)
data=self.func_encode_skip( 1, 0, 0, 0, 0, sel, 0, 0, 0, 1, 0, 0, 0)
self.x393_mem.axi_write_single_w(cmd_addr, data, verbose)
cmd_addr += 1
# Turn off read pattern mode
......@@ -995,6 +997,55 @@ class X393PIOSequences(object):
print ("WLEV lanes ratios: %f %f, non 0x00/0x01 bytes: %f"%(rslt[0],rslt[1],rslt[2]))
return rslt
def read_levelling(self,
nrep,
sel=1, # 0 - early, 1 - late read command (shift by a SDCLK period), -1 - use current sequence
quiet=1):
"""
Read and process data in 'read patter' mode
refresh may be off, delays: cmda_edelay, dq_idelay, dqs_idelauy should be set
<nrep> number of times pattern burst is read (8-bursts), actually will be read nrep+3, nut the first/last will be discarded
<sel> 0 - early, 1 - late read command (shift by a SDCLK period)
"""
if sel>=0:
self.set_read_pattern(nrep+3,sel) # do not use first/last pair of the 32 bit words
buf= self.read_pattern((4*(nrep+1)+2), # num,
(0,1)[quiet<1], # show_rslt,
1) # Wait for operation to complete
buf= buf[4:] # discard first 4*32-bit words
if quiet <1:
hbuf=[]
for dd in buf:
hbuf.append(hex(dd))
print(hbuf)
# with "good" data each word in buf should be 0xff00ff00
for d in buf:
if d !=0xffffffff:
break
else: # Maybe it is not needed with correct cmda_odelay ?
if quiet<3:
print ("Got a block of bad data during read pattern")
return None
data=[0]*32 # for each bit - even, then for all - odd
for w in range (4 * nrep):
lane=w%2
for wb in range(32):
g=(wb/8)%2
b=wb%8+lane*8+16*g
if (buf[w+2] & (1<<wb) != 0):
data[b]+=1
scale=1.0/(4*nrep)
for i in range(32):
data[i]*=scale
if quiet < 2:
for i in range(32):
print("%.2f"%data[i],end=" ")
print()
return data
def restart_ddr3(self,
wait_complete=True,
quiet=1):
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment