[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Xen-devel] "tcp: refine TSO autosizing" causes performance regression on Xen
- To: George Dunlap <george.dunlap@xxxxxxxxxxxxx>
- From: Eric Dumazet <eric.dumazet@xxxxxxxxx>
- Date: Wed, 15 Apr 2015 10:41:28 -0700
- Cc: Jonathan Davies <Jonathan.Davies@xxxxxxxxxx>, "xen-devel@xxxxxxxxxxxxxxxxxxx" <xen-devel@xxxxxxxxxxxxxxxxxxx>, Wei Liu <wei.liu2@xxxxxxxxxx>, Ian Campbell <Ian.Campbell@xxxxxxxxxx>, Stefano Stabellini <stefano.stabellini@xxxxxxxxxxxxx>, netdev <netdev@xxxxxxxxxxxxxxx>, Linux Kernel Mailing List <linux-kernel@xxxxxxxxxxxxxxx>, Eric Dumazet <edumazet@xxxxxxxxxx>, Paul Durrant <paul.durrant@xxxxxxxxxx>, linux-arm-kernel@xxxxxxxxxxxxxxxxxxx, Felipe Franciosi <felipe.franciosi@xxxxxxxxxx>, Christoffer Dall <christoffer.dall@xxxxxxxxxx>, David Vrabel <david.vrabel@xxxxxxxxxx>
- Delivery-date: Wed, 15 Apr 2015 17:41:38 +0000
- List-id: Xen developer discussion <xen-devel.lists.xen.org>
On Wed, 2015-04-15 at 18:23 +0100, George Dunlap wrote:
> Which means that max(2*skb->truesize, sk->sk_pacing_rate >>10) is
> *already* larger for Xen; that calculation mentioned in the comment is
> *already* doing the right thing.
Sigh.
1ms of traffic at 40Gbit is 5 MBytes
The reason for the cap to /proc/sys/net/ipv4/tcp_limit_output_bytes is
to provide the limitation of ~2 TSO packets, which _also_ is documented.
Without this limitation, 5 MBytes could translate to : Fill the queue,
do not limit.
If a particular driver needs to extend the limit, fine, document it and
take actions.
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel
|
