Generating the coefficients for a Xilinx CFIR in Matlab

I'm trying to implement a digital down-conversion using the Xilinx DDC core
on an FPGA.  I have a question about generating the CFIR coefficients from
Matlab when you want a decimation rate of 2 (or more) from the CFIR. 
Matlab won't allow you to directly generate the coefficients in xilinx
format using the command 'coewrite(HCfir)' because it is a multi-rate
filter object (not sure what the problem is with that).  Let's say I tell
the Xilinx Coregen that I want a 20-tap symmetric CFIR with R=2.  How are
the coefficients in the .coe file ordered?  Are the first 10 coefficients
for the higher sampling rate input and the second 10 for after decimation?


Also is there a work around to get Matlab to just generate the
coefficients for me using coewrite?