[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Wed, 15 Jun 2016 01:40:35 +0000
From: "Brown, Aaron F" <aaron.f.brown@...el.com>
To: Andrew Lunn <andrew@...n.ch>,
"Kirsher, Jeffrey T" <jeffrey.t.kirsher@...el.com>,
David Miller <davem@...emloft.net>
CC: netdev <netdev@...r.kernel.org>,
"intel-wired-lan@...ts.osuosl.org" <intel-wired-lan@...ts.osuosl.org>
Subject: RE: [Intel-wired-lan] [PATCHv2 net-next] net: igb: Only dma sync
frame length
> From: Intel-wired-lan [mailto:intel-wired-lan-bounces@...ts.osuosl.org] On
> Behalf Of Andrew Lunn
> Sent: Friday, June 3, 2016 2:03 PM
> To: Kirsher, Jeffrey T <jeffrey.t.kirsher@...el.com>; David Miller
> <davem@...emloft.net>
> Cc: netdev <netdev@...r.kernel.org>; intel-wired-lan@...ts.osuosl.org;
> Andrew Lunn <andrew@...n.ch>
> Subject: [Intel-wired-lan] [PATCHv2 net-next] net: igb: Only dma sync frame
> length
>
> On some platforms, syncing a buffer for DMA is expensive. Rather than
> sync the whole 2K receive buffer, only synchronise the length of the
> frame, which will typically be the MTU, or a much smaller TCP ACK.
>
> For an IMX6Q, this gives around 6% increased TCP receive performance,
> which is cache operations bound and reduces CPU load for TCP transmit.
>
> Signed-off-by: Andrew Lunn <andrew@...n.ch>
> ---
> v2:
> Christmas tree the local variables
> Pass size into igb_add_rx_frag() rather than repeating the endiness swap.
> ---
> drivers/net/ethernet/intel/igb/igb_main.c | 7 ++++---
> 1 file changed, 4 insertions(+), 3 deletions(-)
Tested-by: Aaron Brown <aaron.f.brown@...el.com>
Powered by blists - more mailing lists