On 2 February 2017 at 13:40, Joe Stringer <j...@ovn.org> wrote:
> On 1 February 2017 at 11:06, Joe Stringer <j...@ovn.org> wrote:
>> On 31 January 2017 at 16:57, Joe Stringer <j...@ovn.org> wrote:
>>> /* Returns true if a variable length meta-flow field 'mff' is not mapped in
>>> * the 'vl_mff_map'. */
>>> bool
>>> -mf_vl_mff_not_mapped(const struct mf_field *mff,
>>> - const struct vl_mff_map *vl_mff_map)
>>> +mf_vl_mff_mapped(const struct mf_field *mff, const struct vl_mff_map *map)
>>> {
>>> - if (mff && vl_mff_map) {
>>> - if (mff->variable_len && !mff->mapped) {
>>> - return true;
>>> - }
>>> - }
>>> -
>>> - return false;
>>> + return !(map && mff && mff->variable_len && !mff->mapped);
>>> }
>>
>> Yi-Hung pointed out offline that this reversal doesn't quite sit right
>> logically; this function is searching for a specific set of invalid
>> conditions, where there is a vl_mff_map, and the field is
>> variable-length, and it's not mapped. It's misleading to have all of
>> this covered by a function named "...mapped()".
>>
>> I suggest we retain the original logic but rename the function to
>> something like mf_vl_mff_invalid().
>
> I applied this to master.
Also cherry-picked to branch-2.7.
_______________________________________________
dev mailing list
d...@openvswitch.org
https://mail.openvswitch.org/mailman/listinfo/ovs-dev
