Forum Discussion
Hi KhaiY,
I discussed the issue with a colleague of mine and I have found something interesting afterwards.
The script sdc_util.tcl located in the synth folder of the altera_lvds_core which is a subfolder of lvds_rx and lvds_tx and the sdc file, lvds_rx_altera_lvds_core20_191_5u346cy.sdc, expect that there is a signal called ..|arch_inst|default_lvds_clock_tree.lvds_clock_tree_inst|lvdsfclk_in within the lvds rx instance. However, I can see, for instance, in the technology map viewer that the instance, lvds_clock_tree_inst, is only part of the lvds_tx core and the lvds_rx instance only uses the lvdsfclk_out signal. In other words, the lvds_rx has no lvds_clock_tree_inst. We can see this fact in the list of atoms in the technology map viewer as well. Hence, there exists no node called ...|lvds_rx|lvds_0|core|arch_inst|default_lvds_clock_tree.lvds_clock_tree_inst|lvdsfclk_in, which is the node causing the error.
Setting lvds_clock_tree_inst_name within lvds_rx_altera_lvds_core20_191_5u346cy.sdc removed some warnings. However, the sdc_util.tcl is very generic and the part regarding lvds_rx is also executed in the sdc_util.tcl of the lvds_tx. Therefore, I was not able to assign fclk_name to the correct location, since this would break other iterations of the loop starting in line 334. Moreover, this would not be the correct way anyway, since these are auto-generated scripts.
Do you have any solution for us to assign the correct location of the lvds_clock_tree_inst for the lvds_rx instance?
Thank you.
Michael