On Wed, Nov 06, 2024 at 09:34:38AM +0100, Florian Westphal wrote: > Pablo Neira Ayuso <pablo@xxxxxxxxxxxxx> wrote: > > > Can you clarify? Do you mean skb_tstamp() vs ktime_get_real_ns() > > > or tstamp sampling in general? > > > > I am referring to ktime_get_real_ns(), I remember to have measured > > 25%-30% performance drop when this is used, but I have not refreshed > > those numbers for long time. > > > > As for skb_tstamp(), I have to dig in the cost of it. > > Its not about the cost, its about the sampling method. > If skb has the rx timestamp, then the event will reflect the skb > creation/rx time, not the "event time". Did that make sense? I think ktime_get_real_ns() needs to be used to get the "event time", I am afraid skb_tstamp() is not useful.
