On Fri, Dec 07, 2012 at 03:48:22PM +0200, Jarno Rajahalme wrote: > > Signed-off-by: Jarno Rajahalme <jarno.rajaha...@nsn.com>
This looks pretty good. At first examination, the only issue I have with it is that it puts special cases for different versions of otherwise similar messages in the code that needs to handle those messages. The strategy that I prefer is to define a pair of functions (generally in ofp-util.[ch]) that converts between the representation of a message in particular protocols and a neutral in-memory representation. ofp-util.c[h] already has numerous examples that you can use as models. Would you mind recasting your patch in this form? Thanks, Ben. _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
