Thread (45 messages) 45 messages, 10 authors, 2013-05-23

Re: [PATCH v3 net-next 1/4] net: implement support for low latency socket polling

flat view

From: Eric Dumazet <hidden>
Date: 2013-05-21 19:30:00
Also in: lkml

On Tue, 2013-05-21 at 22:25 +0300, Eliezer Tamir wrote:
On 21/05/2013 20:51, Eric Dumazet wrote:
quoted
On Tue, 2013-05-21 at 10:48 -0700, Eric Dumazet wrote:
quoted
We do not keep skb->dev information once a packet leaves the rcu
protected region.

Once packet is queued to tcp input queues, skb->dev is NULL.
This is done in tcp_v4_rcv() & tcp_v6_rcv()
So if we move calling sk_mark_ll() into the rcu protected region,
We would not need the gen id in the skb.
We could save skb->dev and skb->queue_mapping along with the generation 
id in the socket.

then we don't need to change the skb struct at all.
skb_mark_ll() can go away.
we also call sk_mark_ll in one central place in net_protocol->handler 
instead of all over the place.

or am I missing something?

You cannot keep a pointer to some object without taking a reference on
the object.

Thats why I suggested to use an napi identifier instead, this is safe
and cheap.




------------------------------------------------------------------------------
Try New Relic Now & We'll Send You this Cool Shirt
New Relic is the only SaaS-based application performance monitoring service 
that delivers powerful full stack analytics. Optimize and monitor your
browser, app, & servers with just a few lines of code. Try New Relic
and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_may
_______________________________________________
E1000-devel mailing list
E1000-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/e1000-devel
To learn more about Intel&#174; Ethernet, visit http://communities.intel.com/community/wired
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help