> On Dec 1, 2016, at 3:44 PM, Ben Pfaff <b...@ovn.org> wrote:
> On Thu, Dec 01, 2016 at 01:09:12PM -0800, Jarno Rajahalme wrote:
>>> If I may be permitted to nit-pick, the name "modify_forward_counts" took
>>> me a bit of thinking to properly understand. Maybe "modify_keep_stats"
>>> would be easier for me to understand at first glance.
>> “stats” include the last used timestamp, which is treated
>> independently of the byte and packet counts. How about
Pushed to master and branch-2.6.
Backported the earlier used-timestamp fix to branch-2.5, and squashed in a fix
for that patch, since it was the one introducing the reset_counts regression,
and also squashed in the new test case to verify behavior on branch-2.5.
Branch-2.5 is OK apart from a difference in reported packet size (54 vs. 60
bytes). Since that difference is unrelated, I modified the test case and pushed
the result to branch-2.5.
dev mailing list