I answered a similar question in this thread last week. Set the diode to have the area set to pPar("l")*pPar("w"). The LVS netlist (CDL, I assume) would also have the expression (l*w) but that shouldn't be a problem for any LVS tool to be able to resolve during LVS.
In reply to Andrew Beckett:
Thanks for your help!
We want to replace the function expression with real numbers because we have perl program to transfer the cdl output to Calibre LVS netlist format.
We use PDK callback(from the other cell) to solve some function expression problem.
Unfortunately, there is no area callback function in the PDK.
Is it possible to write a simple callback to solve the problem?
cellId = ddGetObj("library" "cellname" "cellview")
cdfgData = cdfGetInstCDF(cellId)
cdfgData->area->value=sprintf(s "%g" (l*w))
In reply to geeknerd:
But Calibre is perfectly capable of evaluating the expression - so why would you do this?
A callback cannot solve this problem. If you have a block (say "myblock") which accepts CDF parameters w and l, and within myblock you have an instance of a diode which needs an area, as I said, the diode should allow the area to be set as pPar("w")*pPar("l").
Imagine instead that you wanted to do this with a callback. You have two instances of myblock, one with w=1u,l=1u and one with w=2u,l=3u. If you (somehow) set the area on the diode instance inside myblock to be w*l for one instance, it will be wrong for the other. Potentially you could have a pseudo-flattening netlister which creates a separate subckt for each variant, but that's a lot of work (I wrote one of these many years ago (about 20) when Dracula's expression evaluation didn't really work yet), and complete overkill when the Verification tool can evaluate the expression. Essentially you need the expression to be evalated as the design is elaborated in the tool that reads the netlist - be it the simulator or the LVS tool.
It sounds as if you must have the incorrect switch or stop view list setup. You need to ensure that the switch list contains the name of the view to switch into, and that it doesn't have a stop view listed before the schematic view.
If you still have a problem, please contact customer support so that somebody can look at your data to understand what is wrong with your set up. This is pretty straightforward and not unusual though, so I'm sure it's just something wrong in how you've told it to netlist.