On Wed, Nov 7, 2012 at 11:13 AM, Pravin Shelar <[email protected]> wrote:
> > > > On Wed, Nov 7, 2012 at 8:58 AM, Jesse Gross <[email protected]> wrote: > >> On Tue, Nov 6, 2012 at 8:04 PM, Pravin Shelar <[email protected]> wrote: >> >>> >>> >>> >>> On Tue, Nov 6, 2012 at 6:32 PM, Jesse Gross <[email protected]> wrote: >>> >>>> If a key used to find a port then we mark it as being not present >>>> (previously this was accomplished by setting it to zero). However, >>>> we also still pass the key to userspace which could lead to an >>>> inconsistent state. This also zeros out the key as before in order >>>> to make our reported results consistent. >>>> >>>> >>> I am not sure why is it inconsistent, zeroing key can not mean it is not >>> present since zero is valid key. only key_present flag can be used for >>> check and I think it is set correctly. >>> >> >> It's possible to have a key that is present and zero but it's illogical >> to have a value for a key that nominally doesn't exist. >> > I think key_present flags that matters and value of key not. key value > least can be used for debugging as we getting it for free. > > either ways it does not matter much. > I agree that it doesn't really matter all that much and the problem will go away completely once we have finished the tunneling work. I'm going to keep the patch as it is though since I think as we start to bring the different pieces together I think it is good to have only conditions that should arise "normally" (i.e. it is more similar to the direction we are going). Thanks for the reviews Kyle and Pravin; I'll push this series shortly.
_______________________________________________ dev mailing list [email protected] http://openvswitch.org/mailman/listinfo/dev
