blob: 2b69c9eadfaa8a36bb25fd8330f9edef8ae48981 (
plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
|
#
# Copyright 2018 Ettus Research, A National Instruments Company
# SPDX-License-Identifier: LGPL-3.0
#
# 20 MHz White Rabbit reference
create_clock -name wr_refclk -period 50.000 [get_ports WB_20MHZ_P]
# 62.5M TXOUTCLK from the MGT that does most of the heavy lifting
create_clock -name wr_txoutclk -period 16.000 \
[get_pins -hierarchical -filter \
{NAME=~sfp_wrapper_0/*/cmp_xwrc_platform/gen_phy_kintex7.cmp_gtx/U_GTX_INST/gtxe2_i/TXOUTCLK}]
# 62.5M RXOUTCLK from the MGT that doesn't do as much as the TXOUTCLK
create_clock -name wr_rxoutclk -period 16.000 \
[get_pins -hierarchical -filter \
{NAME=~sfp_wrapper_0/*/cmp_xwrc_platform/gen_phy_kintex7.cmp_gtx/U_GTX_INST/gtxe2_i/RXOUTCLK}]
# Rename the WR 62.5M system clock
create_generated_clock -name wr_sysclk \
[get_pins -hierarchical -filter {NAME=~sfp_wrapper_0/*/cmp_xwrc_platform/gen_default_plls.gen_kintex7_default_plls.cmp_sys_clk_pll/CLKOUT0}]
# Rename the WR ~62.5M DMTD clock
create_generated_clock -name wr_dmtdclk \
[get_pins -hierarchical -filter {NAME=~sfp_wrapper_0/*/cmp_xwrc_platform/gen_default_plls.gen_kintex7_default_plls.cmp_dmtd_clk_pll/CLKOUT0}]
# Clock Interaction Matrix : ------------------------------------------------------------
# Everything is basically async to everything else, except for the net_clk group.
# Blanks mean there are no paths there (in the WR v4.2 build).
#
# The wr_sysclk (62.5 MHz) is derived from the 125 MHz MGT Reference clock (NETCLK).
# On the other hand, the wr_dmtdclk, also a 62.5 MHz clock, is derived from the 20 MHz
# WR VCXO. This two clock groups could be considered asynchronous due to its different
# source and different MMCM primitive (which unrelates them even more).
#
# | wr_RXOUTCLK | wr_TXOUTCLK | wr_dmtdclk | wr_sysclk | net_clk
# wr_RXOUTCLK | timed | async | async | async |
# wr_TXOUTCLK | async | timed | async | async |
# wr_dmtdclk | | | timed | async |
# wr_sysclk | async | async | async | timed | timed
# net_clk | async | | | timed | timed
set_clock_groups -asynchronous -group [get_clocks wr_rxoutclk] \
-group [get_clocks wr_txoutclk] \
-group [get_clocks wr_dmtdclk] \
-group [get_clocks net_clk]
set_clock_groups -asynchronous -group [get_clocks wr_sysclk] \
-group [get_clocks wr_txoutclk] \
-group [get_clocks wr_rxoutclk] \
-group [get_clocks wr_dmtdclk]
# Safe paths : --------------------------------------------------------------------------
# There is a mix of safe and unknown clock interactions between wr_sysclk and wr_dmtdclk.
# The safe crossings refer to those using a synchronizer structure; and to avoid
# failing timing on those paths, we ignore them (i.e. false paths), and allow the tools
# to time the rest.
#
# The WR IP uses easily identifiable single-bit synchronizers to transfer strobes from
# the wr_dmtdclk to the wr_sysclk domain.
#
# Commenting out command since the asynchronous groups were updated and paths ignored.
# set_false_path -from [get_clocks wr_dmtdclk] \
# -to [get_pins -hierarchical -filter {NAME =~ */sync_posedge.sync0_reg/D}]
#
#
# The tags generated by the DDMTD (wr_dmtdclk domain) are deglitched before being output
# to the upstream logic in the wr_sysclk domain. Thus, a safe multi-bit crossing is
# expected between these two domains, and we tell the tools to ignore those paths.
# See usrp3/lib/white_rabbit/wr_cores_v4_2/modules/timing/dmtd_with_deglitcher.vhd for
# implementation details.
#
# Commenting out command since the asynchronous groups were updated and paths ignored.
# set_false_path -from [get_clocks wr_dmtdclk] \
# -to [get_pins -hierarchical -filter {NAME =~ */U_Wrapped_Softpll/tags_masked_reg*/D}]
|
