Date:	Thu, 26 Feb 2015 20:22:40 -0500
From:	Neil Horman <nhorman@...driver.com>
To:	Simon Horman <simon.horman@...ronome.com>
Cc:	John Fastabend <john.r.fastabend@...el.com>,
	Thomas Graf <tgraf@...g.ch>, Jiri Pirko <jiri@...nulli.us>,
	netdev@...r.kernel.org, davem@...emloft.net, andy@...yhouse.net,
	dborkman@...hat.com, ogerlitz@...lanox.com, jesse@...ira.com,
	jpettit@...ira.com, joestringer@...ira.com, jhs@...atatu.com,
	sfeldma@...il.com, f.fainelli@...il.com, roopa@...ulusnetworks.com,
	linville@...driver.com, shrijeet@...il.com,
	gospo@...ulusnetworks.com, bcrl@...ck.org
Subject: Re: Flows! Offload them.

On Fri, Feb 27, 2015 at 06:52:58AM +0900, Simon Horman wrote:
> On Thu, Feb 26, 2015 at 03:16:35PM -0500, Neil Horman wrote:
> > On Thu, Feb 26, 2015 at 07:23:36AM -0800, John Fastabend wrote:
> > > On 02/26/2015 05:33 AM, Thomas Graf wrote:
> > > > On 02/26/15 at 10:16am, Jiri Pirko wrote:
> > > >> Well, on netdev01, I believe that a consensus was reached that for every
> > > >> switch offloaded functionality there has to be an implementation in
> > > >> kernel.
> > > > 
> > > > Agreed. This should not prevent the policy being driven from user
> > > > space though.
> > > > 
> > > >> What John's Flow API originally did was to provide a way to
> > > >> configure hardware independently of kernel. So the right way is to
> > > >> configure kernel and, if hw allows it, to offload the configuration to hw.
> > > >>
> > > >> In this case, seems to me logical to offload from one place, that being
> > > >> TC. The reason is, as I stated above, the possible conversion from OVS
> > > >> datapath to TC.
> > > > 
> > > > Offloading of TC definitely makes a lot of sense. I think that even in
> > > > that case you will already encounter independent configuration of
> > > > hardware and kernel. Example: The hardware provides a fixed, generic
> > > > function to push up to n bytes onto a packet. This hardware function
> > > > could be used to implement TC actions "push_vlan", "push_vxlan",
> > > > "push_mpls". You would you would likely agree that TC should make use
> > > > of such a function even if the hardware version is different from the
> > > > software version. So I don't think we'll have a 1:1 mapping for all
> > > > configurations, regardless of whether the how is decided in kernel or
> > > > user space.
> > > 
> > > Just to expand slightly on this. I don't think you can get to a 1:1
> > > mapping here. One reason is hardware typically has a TCAM and limited
> > > size. So you need a _policy_ to determine when to push rules into the
> > > hardware. The kernel doesn't know when to do this and I don't believe
> > > its the kernel's place to start enforcing policy like this. One thing I likely
> > > need to do is get some more "worlds" in rocker so we aren't stuck only
> > > thinking about the infinite size OF_DPA world. The OF_DPA world is only
> > > one world and not a terribly flexible one at that when compared with the
> > > NPU folk. So minimally you need a flag to indicate rules go into hardware
> > > vs software.
> > > 
> > > That said I think the bigger mismatch between software and hardware is
> > > you program it differently because the data structures are different. Maybe
> > > a u32 example would help. For parsing with u32 you might build a parse
> > > graph with a root and some leaf nodes. In hardware you want to collapse
> > > this down onto the hardware. I argue this is not a kernel task because
> > > there are lots of ways to do this and there are trade-offs made with
> > > respect to space and performance and which table to use when it could be
> > > handled by a set of tables. Another example is a virtual switch possibly
> > > OVS but we have others. The software does some "unmasking" (there term)
> > > before sending the rules into the software dataplane cache. Basically this
> > > means we can ignore priority in the hash lookup. However this is not how you
> > > would optimally use hardware. Maybe I should do another write up with
> > > some more concrete examples.
> > > 
> > > There are also lots of use cases to _not_ have hardware and software in
> > > sync. A flag allows this.
> > > 
> > > My only point is I think we need to allow users to optimally use there
> > > hardware either via 'tc' or my previous 'flow' tool. Actually in my
> > > opinion I still think its best to have both interfaces.
> > > 
> > > I'll go get some coffee now and hopefully that is somewhat clear.
> > 
> > 
> > I've been thinking about the policy apect of this, and the more I think
> > about it, the more I wonder if not allowing some sort of common policy in
> > the kernel is really the right thing to do here.  I know thats somewhat
> > blasphemous, but this isn't really administrative poilcy that we're
> > talking about, at least not 100%.  Its more of a behavioral profile that
> > we're trying to enforce.  That may be splitting hairs, but I think theres
> > precidence for the latter.  That is to say, we configure qdiscs to limit
> > traffic flow to certain rates, and configure policies which drop traffic
> > that violates it (which includes random discard, which is the antithesis
> > of deterministic policy).  I'm not sure I see this as any different,
> > espcially if we limit its scope.  That is to say, why couldn't we allow
> > the kernel to program a predetermined set of policies that the admin can
> > set (i.e. offload routing to a hardware cache of X size with an lru
> > victimization).  If other well defined policies make sense, we can add
> > them and exposes options via iproute2 or some such to set them.  For the
> > use case where such pre-packaged policies don't make sense, we have
> > things like the flow api to offer users who want to be able to control
> > their hardware in a more fine grained approach.
> 
> In general I agree that it makes sense to have have sane offload policy
> in the kernel and provide a mechanism to override that. Things that already
> work should continue to work: just faster or with fewer CPU cycles consumed.
> 
Yes, exactly that, for the general traditional networking use case, that is
exactly what we want, to opportunistically move traffic faster with less load on
the cpu.  We don't nominally care what traffic is offloaded, as long as the
hardware does a better job than just software alone.  If we get an occasional
miss and have to do stuff in software, so be it.

> I am, however, not entirely convinced that it is always possible to
> implement such a sane default policy that is worth the code complexity -
> I'm thinking in particular of Open vSwitch where management of flows is
> already in user-space.
So, this is a case in which I think John F.'s low level flow API is more well
suited.  OVS has implemented a user space dataplane that circumvents alot of the
kernel mechanisms for traffic forwarding.  For that sort of application, the
traditional kernel offload "objects" aren't really appropriate.  Instead, OVS
can use the low level flow API to construct its own custom offload pipeline
using whatever rules and policies that it wants.

Of course, using the low level flow API is incompatible with the in-kernel
object offload idea that I'm proposing, but I see the two as able to co-exist,
much like firewalld co-exists with iptables.  You can use both, but you have to
be aware that using the lower layer interface might break the others higher
level oeprations.  And if that happens, its on you to manage it.

Best
Neil

> 
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Powered by blists - more mailing lists