> On Mar 17, 2015, at 1:39 PM, Ben Pfaff <b...@nicira.com> wrote: > > On Tue, Mar 17, 2015 at 01:33:40PM -0700, Jarno Rajahalme wrote: >> >>> On Mar 17, 2015, at 1:20 PM, Ben Pfaff <b...@nicira.com> wrote: >>> >>> On Fri, Mar 13, 2015 at 04:51:59PM -0700, Jarno Rajahalme wrote: >>>> Signed-off-by: Jarno Rajahalme <jrajaha...@nicira.com> >>> >>> It looks like ->use_recirc and ->recirc in xlate_ctx are now only an >>> interface from output_normal() to compose_output_action__(). Can we >>> instead make a "struct xlate_recirc *" a parameter passed along that >>> call chain, with a NULL value coresponding to use_recirc == false? >>> >> >> I did consider this, and agree that removing this from xlate_ctx >> makes it cleaner, so, will do :-) > > OK. Hearing that you considered it makes me wonder whether there are > reasons why it's a bad idea?
I should have said “briefly considered”, as in, "the idea crossed my mind, but I did not follow through" :-) The only counter argument I had was that it may complicate the function prototypes, and as you earlier commented on the record parameter to xlate_actions(), I decided to not follow through with the idea. I have now done the change, and there were only two affected functions, compose_output_action() and compose_output_action__(), which both are static functions, so no uglier APIs because of this. Jarno _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
