<div dir="ltr">I thought the discussion was only about GSO/TSO. Also, isn't GRO/LRO incompatible with routing? Anyway, I was just supporting your interpretation of what Eric potentially means.<div><br></div><div>/Jonas</div></div><div class="gmail_extra"><br><div class="gmail_quote">On Thu, Apr 26, 2018 at 10:55 AM, Toke Høiland-Jørgensen <span dir="ltr"><<a href="mailto:toke@toke.dk" target="_blank">toke@toke.dk</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><span class="">Jonas Mårtensson <<a href="mailto:martensson.jonas@gmail.com">martensson.jonas@gmail.com</a>> writes:<br>
<br>
> "I *think* that what Eric means is that the GSO logic should automatically<br>
> size the GSO superpackets so the latency cost is negligible for the actual<br>
> link rate."<br>
><br>
> Something like this?<br>
><br>
> <a href="https://lwn.net/Articles/564979/" rel="noreferrer" target="_blank">https://lwn.net/Articles/<wbr>564979/</a><br>
><br>
> <a href="https://lwn.net/Articles/564978/" rel="noreferrer" target="_blank">https://lwn.net/Articles/<wbr>564978/</a><br>
<br>
</span>Yeah, that's for TCP on the local host. I'm not sure how things work for<br>
GRO on a router (which is more relevant for the CAKE use case).<br>
<span class="HOEnZb"><font color="#888888"><br>
-Toke<br>
</font></span></blockquote></div><br></div>