lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date:   Tue, 26 Feb 2019 09:14:31 +0000
From:   Ioana Ciornei <ioana.ciornei@....com>
To:     Jesper Dangaard Brouer <brouer@...hat.com>,
        Ioana Ciocoi Radulescu <ruxandra.radulescu@....com>
CC:     Ilias Apalodimas <ilias.apalodimas@...aro.org>,
        "netdev@...r.kernel.org" <netdev@...r.kernel.org>,
        "davem@...emloft.net" <davem@...emloft.net>
Subject: RE: [PATCH net-next 1/4] dpaa2-eth: Use a single page per Rx buffer


> Subject: RE: [PATCH net-next 1/4] dpaa2-eth: Use a single page per Rx buffer
> 
> > Subject: Re: [PATCH net-next 1/4] dpaa2-eth: Use a single page per Rx
> > buffer
> >
> >
> > On Wed, 6 Feb 2019 15:36:33 +0000 Ioana Ciocoi Radulescu
> > <ruxandra.radulescu@....com> wrote:
> >
> > > > From: Ilias Apalodimas <ilias.apalodimas@...aro.org>
> > > >
> > > > Can you share any results on XDP (XDP_DROP is usually useful for
> > > > the hardware capabilities).
> > >
> > > XDP numbers are pretty much the same as before this patch:
> > >
> > > On a LS2088A with A72 cores @2GHz (numbers in Mpps):
> > > 				1core		8cores
> > > -------------------------------------------------------------------------
> > > XDP_DROP (no touching data)	5.37		29.6 (linerate)
> > > XDP_DROP (xdp1 sample)	3.14		24.22
> >
> > It is interesting/problematic to see that the cost of touching the
> > data is so high 5.37Mpps -> 3.14Mpps.  The Intel CPUs have solved this
> > in hardware with DDIO, which delivers frame in L3-cache. I have some
> > ideas on how to improve this on ARM (or CPUs without DDIO).  I've
> > previous implemented this as RFC on mlx4 tested on a CPU without DDIO,
> > with great success 10mpps -> 20Mpps (but it was shutdown, as newer
> > Intel HW solved the issue).  The basic idea is to have an array of
> > frames, that you start an L2/L3-prefetch on, before going "back" and
> > process them for XDP or netstack. (p.s. this is the same DPDK does)
> 
> Thanks for the hint. We are currently investigating what our options are.
> We'll come back with a more detailed answer once we figure out.
> 

Hi,

We have enabled cache stashing of frame data in our solution and sent the patches upstream: https://lkml.org/lkml/2019/2/23/41.

Some results with the patches applied on top of the net-next tree are below.
LS2088A with A72 cores @2GHz (numbers in Mpps):

 			                		1core		8cores
 --------------------------------------------------------------------------------------------------------
 XDP_DROP (no touching data)			5.37		29.6 (linerate)
 XDP_DROP (xdp1 sample, no stashing)		3.28		24.2
 XDP_DROP (xdp1 sample, stashing)		4.64		26.85

Thanks a lot for your hint.

Ioana C

Powered by blists - more mailing lists