<DIV>i changed the code to pass packets if they don't need to be delayed. </DIV>
<DIV>could there be some other issues if packets are just passed through?</DIV>
<DIV> </DIV>
<DIV>thanks. <BR><BR><B><I>Stephen Hemminger <shemminger@osdl.org></I></B> wrote:</DIV>
<BLOCKQUOTE class="replbq" DEFANGED_style="PADDING-LEFT: 5px; MARGIN-LEFT: 5px; BORDER-LEFT: #1010ff 2px solid">><BR>><BR>> think this is probably because netem sticks the<BR>>incoming pkt into its delay queue with no delay, and<BR>>then pulls it out when it services this queue x<BR>>milliseconds later, where x is the resolution of<BR>>jiffies (1ms on x86 by default).<BR>><BR><BR>Yes, that is correct. It would be possible to bypass that but what would<BR>the point be? The experimental hack to allow re-order does call packets<BR>with <1ms delay to pass right through, but that was more of a fall-out<BR>from the recoding than an explicit design choice. I am easy on this<BR>and perfectly willing to implement what ever seems to make sense.<BR></BLOCKQUOTE><p>
<hr size=1>Do you Yahoo!?<br>
Yahoo! Search presents - <a href="http://us.rd.yahoo.com/evt=30648/*http://movies.yahoo.com/movies/feature/jibjabinaugural.html">Jib Jab's 'Second Term'</a>