Hi
Raghavendra, <div><br></div><div>"
/proc/sys/net/core/bpf_jit_<u></u>enable " is not present on machine. Seems kernel need to be configured for these. <div><div>Btw CPU scaling issue is resolved. Packet drop was because all interrupt handling was happening on core where packets were getting drop. Tuning system for interrupt handling using smp_affinity resolved the issue. </div>
<div><br></div><div>Thanks</div><div>Mukesh<br><br><div class="gmail_quote">On 22 December 2011 19:18, Raghavendra D Prabhu <span dir="ltr"><<a href="mailto:raghu.prabhu13@gmail.com">raghu.prabhu13@gmail.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Hi,<br>
<br>
Have you tried enabling -- /proc/sys/net/core/bpf_jit_<u></u>enable --<br>
provides JIT compiled BPF on 64 bit linux.<div><div></div><div class="h5"><br>
<br>
* On Wed, Dec 21, 2011 at 12:08:04AM +0530, Mukesh Yadav <<a href="mailto:mukesh.fkd@gmail.com" target="_blank">mukesh.fkd@gmail.com</a>> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Hi Guys,<br>
<br>
I am working on user level code which uses BPF filter.<br>
Solution has two threads and BPF filter for incoming packets to raw socket<br>
dedicated to each thread.<br>
Intent is to divide ingress traffic between two thread based on whether<br>
dest IP is even or odd.<br>
<br>
BPF filters used are:<br>
Even IP filter :-> tcpdump -i interface 'ether dst <dest mac> && ip[19] &<br>
0x01 = 0'<br>
Odd IP filter :-> tcpdump -i interface 'ether dst <dest mac> && ip[19] &<br>
0x01 = 1'<br>
<br>
There is a packet loss for odd IP thread even when CPU is available. Reason<br>
being packet drop due recv buffer full.<br>
Same amount of traffic is well handled by even IP thread(user code being<br>
same in both), scaling CPU to full 100%.<br>
<br>
In odd IP thread, If filter is changed to "tcpdump -i interface 'ether dst<br>
<dest mac> && ip proto 17' ", all goes fine.<br>
Also CPU usage at kernel drops from 50% to 4 % for a particular amount of<br>
traffic.<br>
<br>
Would appreciate any I/ps for reducing load at kernel for packet<br>
filtering(Odd/Even IP distribution)<br>
<br>
<br>
Cheers<br>
Mukesh<br>
</blockquote>
<br>
</div></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
______________________________<u></u>_________________<br>
Kernelnewbies mailing list<br>
<a href="mailto:Kernelnewbies@kernelnewbies.org" target="_blank">Kernelnewbies@kernelnewbies.<u></u>org</a><br>
<a href="http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies" target="_blank">http://lists.kernelnewbies.<u></u>org/mailman/listinfo/<u></u>kernelnewbies</a><br>
</blockquote>
<br>
--------------------------<br>
Raghavendra Prabhu<br>
GPG Id : 0xD72BE977<br>
Fingerprint: B93F EBCB 8E05 7039 CD3C A4B8 A616 DCA1 D72B E977<br>
www: <a href="http://wnohang.net" target="_blank">wnohang.net</a><br>
</blockquote></div><br><br clear="all"><div><br></div>-- <br>Cheers<br>Mukesh Yadav<br>Senior Software Engineer<br>Plot6 Sector 18<br>Aricent <br>Gurgaon<br>Mob: 9953024169<br>
</div></div></div>