<div dir="ltr"><div><div>Hi,<br><br></div>I was tracing the program flow. So basically i figured out that net_rx_action calls process_backlog and i saw that deletion is happening at net_rx_action().<br><br>So i used git blame on that file and saw that this commit performed the deletion. Then when i saw what the commit does, i found that it changed the deletion location from process_backlog to here.<br><br></div>Thanks and regards,<br>Vignesh Radhakrishnan<br></div><div class="gmail_extra"><br><div class="gmail_quote">On Wed, Dec 24, 2014 at 8:22 AM, lx <span dir="ltr"><<a href="mailto:lxgeek@yahoo.com" target="_blank">lxgeek@yahoo.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>
How to find this commit:<br>
##############################<u></u>###<br>
Earlier we had deletion in process_backlog itself, but <a href="https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=d75b1ade567ffab085e8adbbdacf0092d10cd09c" target="_blank">https://git.kernel.org/cgit/<u></u>linux/kernel/git/torvalds/<u></u>linux.git/commit/?id=<u></u>d75b1ade567ffab085e8adbbdacf00<u></u>92d10cd09c</a> has changed it.<br>
##############################<u></u>###<br>
<br>
There are so many commits, and this commit don't show some messages about process_backlog().The messages is:<br>
##############################<u></u>###<br>
commit d75b1ade567ffab085e8adbbdacf00<u></u>92d10cd09c<br>
Author: Eric Dumazet <<a href="mailto:edumazet@google.com" target="_blank">edumazet@google.com</a>><br>
Date: Sun Nov 2 06:19:33 2014 -0800<br>
<br>
net: less interrupt masking in NAPI<br>
<br>
net_rx_action() can mask irqs a single time to transfert sd->poll_list<br>
into a private list, for a very short duration.<br>
<br>
Then, napi_complete() can avoid masking irqs again,<br>
and net_rx_action() only needs to mask irq again in slow path.<br>
<br>
This patch removes 2 couples of irq mask/unmask per typical NAPI run,<br>
more if multiple napi were triggered.<br>
<br>
Note this also allows to give control back to caller (do_softirq())<br>
more often, so that other softirq handlers can be called a bit earlier,<br>
or ksoftirqd can be wakeup earlier under pressure.<br>
<br>
This was developed while testing an alternative to RX interrupt<br>
mitigation to reduce latencies while keeping or improving GRO<br>
aggregation on fast NIC.<br>
<br>
Idea is to test napi->gro_list at the end of a napi->poll() and<br>
reschedule one NAPI poll, but after servicing a full round of<br>
softirqs (timers, TX, rcu, ...). This will be allowed only if softirq<br>
is currently serviced by idle task or ksoftirqd, and resched not needed.<br>
<br>
Signed-off-by: Eric Dumazet <<a href="mailto:edumazet@google.com" target="_blank">edumazet@google.com</a>><br>
Cc: Willem de Bruijn <<a href="mailto:willemb@google.com" target="_blank">willemb@google.com</a>><br>
Signed-off-by: David S. Miller <<a href="mailto:davem@davemloft.net" target="_blank">davem@davemloft.net</a>><br>
##############################<u></u>###<br>
<br>
<br>
Thank you.<br>
</blockquote></div><br><br clear="all"><br>-- <br><div class="gmail_signature"><a href="http://vigneshradhakrishnan.blogspot.com/" target="_blank">http://vigneshradhakrishnan.blogspot.com/</a><br></div>
</div>