* HW accelerated support for setting skb->pkt_type?
@ 2013-12-19 2:29 Florian Fainelli
2013-12-19 2:56 ` Eric Dumazet
2013-12-19 5:16 ` David Miller
0 siblings, 2 replies; 3+ messages in thread
From: Florian Fainelli @ 2013-12-19 2:29 UTC (permalink / raw)
To: netdev, Eric Dumazet, David Miller
Hi all,
I work with two Ethernet controllers which can tell on a per-packet
basis whether this is a broadcast, unicast or multicast, eventually
allowing me to bypass the initial checks done in eth_type_trans().
Is there any other hardware out there which supports that, and is
there any interest in optimizing that operation?
Thanks
--
Florian
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: HW accelerated support for setting skb->pkt_type?
2013-12-19 2:29 HW accelerated support for setting skb->pkt_type? Florian Fainelli
@ 2013-12-19 2:56 ` Eric Dumazet
2013-12-19 5:16 ` David Miller
1 sibling, 0 replies; 3+ messages in thread
From: Eric Dumazet @ 2013-12-19 2:56 UTC (permalink / raw)
To: Florian Fainelli; +Cc: netdev, David Miller
On Wed, 2013-12-18 at 18:29 -0800, Florian Fainelli wrote:
> Hi all,
>
> I work with two Ethernet controllers which can tell on a per-packet
> basis whether this is a broadcast, unicast or multicast, eventually
> allowing me to bypass the initial checks done in eth_type_trans().
>
> Is there any other hardware out there which supports that, and is
> there any interest in optimizing that operation?
Other hardware have this as well, but I doubt it makes any difference,
unless you have an arch with a 16 bytes cache line.
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: HW accelerated support for setting skb->pkt_type?
2013-12-19 2:29 HW accelerated support for setting skb->pkt_type? Florian Fainelli
2013-12-19 2:56 ` Eric Dumazet
@ 2013-12-19 5:16 ` David Miller
1 sibling, 0 replies; 3+ messages in thread
From: David Miller @ 2013-12-19 5:16 UTC (permalink / raw)
To: f.fainelli; +Cc: netdev, eric.dumazet
From: Florian Fainelli <f.fainelli@gmail.com>
Date: Wed, 18 Dec 2013 18:29:37 -0800
> I work with two Ethernet controllers which can tell on a per-packet
> basis whether this is a broadcast, unicast or multicast, eventually
> allowing me to bypass the initial checks done in eth_type_trans().
>
> Is there any other hardware out there which supports that, and is
> there any interest in optimizing that operation?
There is zero benefit from this, since you're going to have to touch
the packet headers anyways and the memory access from that dwarfs
anything else.
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2013-12-19 5:16 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2013-12-19 2:29 HW accelerated support for setting skb->pkt_type? Florian Fainelli
2013-12-19 2:56 ` Eric Dumazet
2013-12-19 5:16 ` David Miller
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.