All of lore.kernel.org
 help / color / mirror / Atom feed
* does intel X520-SR(ixgbe) support RSS on single VLAN?
@ 2011-01-24 14:18 Rui
  2011-01-24 17:09 ` Alexander Duyck
  0 siblings, 1 reply; 8+ messages in thread
From: Rui @ 2011-01-24 14:18 UTC (permalink / raw)
  To: netdev

hi
does intel X520-SR support RSS on single VLAN?

tested with 3 different vlan id and priority packets
What I saw is that all packets were always delivered to the same RxQ.
looks can not get a different RSS index for these packet?
any setting needed?

^ permalink raw reply	[flat|nested] 8+ messages in thread

* Re: does intel X520-SR(ixgbe) support RSS on single VLAN?
  2011-01-24 14:18 does intel X520-SR(ixgbe) support RSS on single VLAN? Rui
@ 2011-01-24 17:09 ` Alexander Duyck
  2011-01-24 18:10   ` Rick Jones
  2011-01-25  2:10   ` Rui
  0 siblings, 2 replies; 8+ messages in thread
From: Alexander Duyck @ 2011-01-24 17:09 UTC (permalink / raw)
  To: Rui; +Cc: netdev, e1000-devel

On 1/24/2011 6:18 AM, Rui wrote:
> hi
> does intel X520-SR support RSS on single VLAN?
>
> tested with 3 different vlan id and priority packets
> What I saw is that all packets were always delivered to the same RxQ.
> looks can not get a different RSS index for these packet?
> any setting needed?
> --
> To unsubscribe from this list: send the line "unsubscribe netdev" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

The X520 should have no problems hashing on a single VLAN tagged frame. 
  However the VLAN will not be a part of the RSS hash.  The  only 
components of the hash are the IPv4/IPv6 source and destination 
addresses, and if the flow is TCP then the port numbers.

I would recommend testing with something like the "netperf -t TCP_CRR" 
test which should open a number of ports and spread traffic out between 
multiple queues.

Thanks,

Alex

^ permalink raw reply	[flat|nested] 8+ messages in thread

* Re: does intel X520-SR(ixgbe) support RSS on single VLAN?
  2011-01-24 17:09 ` Alexander Duyck
@ 2011-01-24 18:10   ` Rick Jones
  2011-01-25  2:10   ` Rui
  1 sibling, 0 replies; 8+ messages in thread
From: Rick Jones @ 2011-01-24 18:10 UTC (permalink / raw)
  To: Alexander Duyck; +Cc: e1000-devel, netdev, Rui

Alexander Duyck wrote:
> I would recommend testing with something like the "netperf -t TCP_CRR" 
> test which should open a number of ports and spread traffic out between 
> multiple queues.

TCP_CRR - Connect Request Response - it will cycle through almost the entire 
port space as it goes, one at a time.  Any one four-tuple will be unlikely to 
have all that many packets - just the SYN exchange, the request/response 
exchange and then the FIN exchange, so unless there is a tool looking at the 
queues with microsecond granularity, it will appear like it is all happening at 
once :)

If you want to see one queue used for "a while" and then another, I would 
suggest a TCP_RR test with the confidence intervals set to say 30 iterations. 
That will exchange packets for the test duration (global -l option) and then the 
next iteration will have a four-tuple that differs in the client port number 
from the previous (the "server" port number remains fixed through the iterations 
of the TCP_RR test.

One can also run TCP_RR tests in turn, one after the other, but that consumes 
port numbers in twos on both sides.  I suppose that these days with port number 
randomization that's OK, but in "the old days" it tended to mean that the 
control and data ports marched in lock-step and one would always be even the 
other odd, which didn't always work that well with hashes...  The use of the 
confidence intervals with the TCP_RR test deals with that by having only the one 
netperf control connection and then successive data connections.

happy benchmarking,

rick jones

There is also always the full specification of the port numbers and IP's for the 
data connection, though it is a bit more cumbersome.

------------------------------------------------------------------------------
Special Offer-- Download ArcSight Logger for FREE (a $49 USD value)!
Finally, a world-class log management solution at an even better price-free!
Download using promo code Free_Logger_4_Dev2Dev. Offer expires 
February 28th, so secure your free ArcSight Logger TODAY! 
http://p.sf.net/sfu/arcsight-sfd2d
_______________________________________________
E1000-devel mailing list
E1000-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/e1000-devel
To learn more about Intel® Ethernet, visit http://communities.intel.com/community/wired

^ permalink raw reply	[flat|nested] 8+ messages in thread

* Re: does intel X520-SR(ixgbe) support RSS on single VLAN?
  2011-01-24 17:09 ` Alexander Duyck
  2011-01-24 18:10   ` Rick Jones
@ 2011-01-25  2:10   ` Rui
  2011-01-25  3:05     ` Ben Hutchings
  1 sibling, 1 reply; 8+ messages in thread
From: Rui @ 2011-01-25  2:10 UTC (permalink / raw)
  To: Alexander Duyck; +Cc: netdev, e1000-devel

On Tue, Jan 25, 2011 at 1:09 AM, Alexander Duyck
<alexander.h.duyck@intel.com> wrote:
> On 1/24/2011 6:18 AM, Rui wrote:
>>
>> hi
>> does intel X520-SR support RSS on single VLAN?
>>
>> tested with 3 different vlan id and priority packets
>> What I saw is that all packets were always delivered to the same RxQ.
>> looks can not get a different RSS index for these packet?
>> any setting needed?
>> --
>> To unsubscribe from this list: send the line "unsubscribe netdev" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>
> The X520 should have no problems hashing on a single VLAN tagged frame.
>  However the VLAN will not be a part of the RSS hash.  The  only components
> of the hash are the IPv4/IPv6 source and destination addresses, and if the
> flow is TCP then the port numbers.
>
hi alexander
I got these information from the intel community:

'I asked our software engineers about your question, and this is what I learned.
You cannot filter by just VLAN or VLAN priority.  The L4 type will
also play a role in the filter and as such you would only be able to
filter TCP, UDP, and SCTP packets that are bound for a VLAN.
The command itself to setup a filter is “ethtool –U ethX flow-type
tcp4 vlan 0x2000 vlan-mask 0xE000 action Y” where X is the correct
index for the interface and Y is the queue you want to route the
traffic to.  This would have to be repeated for udp4 and sctp4.
I hope this will help.
Mark H"

so my question is that the VLAN is PART of the RSS or not? looks the
perfect filter support vlan id ?can the perfect filter support
wildchar,such as: flow-type ANY?
thanks


> I would recommend testing with something like the "netperf -t TCP_CRR" test
> which should open a number of ports and spread traffic out between multiple
> queues.
>
> Thanks,
>
> Alex
>

^ permalink raw reply	[flat|nested] 8+ messages in thread

* Re: does intel X520-SR(ixgbe) support RSS on single VLAN?
  2011-01-25  2:10   ` Rui
@ 2011-01-25  3:05     ` Ben Hutchings
  2011-01-25  7:22       ` Jon Zhou
  2011-01-25  9:03       ` Rui
  0 siblings, 2 replies; 8+ messages in thread
From: Ben Hutchings @ 2011-01-25  3:05 UTC (permalink / raw)
  To: Rui; +Cc: Alexander Duyck, netdev, e1000-devel

On Tue, 2011-01-25 at 10:10 +0800, Rui wrote:
> On Tue, Jan 25, 2011 at 1:09 AM, Alexander Duyck
> <alexander.h.duyck@intel.com> wrote:
> > On 1/24/2011 6:18 AM, Rui wrote:
> >>
> >> hi
> >> does intel X520-SR support RSS on single VLAN?
> >>
> >> tested with 3 different vlan id and priority packets
> >> What I saw is that all packets were always delivered to the same RxQ.
> >> looks can not get a different RSS index for these packet?
> >> any setting needed?
> >> --
> >> To unsubscribe from this list: send the line "unsubscribe netdev" in
> >> the body of a message to majordomo@vger.kernel.org
> >> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> >
> > The X520 should have no problems hashing on a single VLAN tagged frame.
> >  However the VLAN will not be a part of the RSS hash.  The  only components
> > of the hash are the IPv4/IPv6 source and destination addresses, and if the
> > flow is TCP then the port numbers.
> >
> hi alexander
> I got these information from the intel community:
> 
> 'I asked our software engineers about your question, and this is what I learned.
> You cannot filter by just VLAN or VLAN priority.  The L4 type will
> also play a role in the filter and as such you would only be able to
> filter TCP, UDP, and SCTP packets that are bound for a VLAN.
> The command itself to setup a filter is “ethtool –U ethX flow-type
> tcp4 vlan 0x2000 vlan-mask 0xE000 action Y” where X is the correct
> index for the interface and Y is the queue you want to route the
> traffic to.  This would have to be repeated for udp4 and sctp4.
> I hope this will help.
> Mark H"

The mask specifies bits to be ignored, so if you want to filter on the
basis of only the priority bits you should use vlan-mask 0xfff.  Unless
this is another inconsistency I failed to notice...

> so my question is that the VLAN is PART of the RSS or not?

It's not part of any specified Toeplitz hash.  However, some hardware
supports adding the hash (after indirection) to the queue number
specified by a filter.  Currently the ethtool API doesn't have a way to
request that.

> looks the
> perfect filter support vlan id ?can the perfect filter support
> wildchar,such as: flow-type ANY?

It is possible to specify this using flow-type ether, but the ixgbe
driver does not yet support that (and I have no idea whether the
hardware does).

Ben.

-- 
Ben Hutchings, Senior Software Engineer, Solarflare Communications
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.


^ permalink raw reply	[flat|nested] 8+ messages in thread

* RE: does intel X520-SR(ixgbe) support RSS on single VLAN?
  2011-01-25  3:05     ` Ben Hutchings
@ 2011-01-25  7:22       ` Jon Zhou
  2011-01-25  9:03       ` Rui
  1 sibling, 0 replies; 8+ messages in thread
From: Jon Zhou @ 2011-01-25  7:22 UTC (permalink / raw)
  To: Ben Hutchings, Rui; +Cc: Alexander Duyck, netdev, e1000-devel



> -----Original Message-----
> From: Ben Hutchings [mailto:bhutchings@solarflare.com]
> Sent: Tuesday, January 25, 2011 11:06 AM
> To: Rui
> Cc: Alexander Duyck; netdev@vger.kernel.org; e1000-
> devel@lists.sourceforge.net
> Subject: Re: does intel X520-SR(ixgbe) support RSS on single VLAN?
> 
> On Tue, 2011-01-25 at 10:10 +0800, Rui wrote:
> > On Tue, Jan 25, 2011 at 1:09 AM, Alexander Duyck
> > <alexander.h.duyck@intel.com> wrote:
> > > On 1/24/2011 6:18 AM, Rui wrote:
> > >>
> > >> hi
> > >> does intel X520-SR support RSS on single VLAN?
> > >>
> > >> tested with 3 different vlan id and priority packets
> > >> What I saw is that all packets were always delivered to the same
> RxQ.
> > >> looks can not get a different RSS index for these packet?
> > >> any setting needed?
> > >> --
> > >> To unsubscribe from this list: send the line "unsubscribe netdev"
> in
> > >> the body of a message to majordomo@vger.kernel.org
> > >> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> > >
> > > The X520 should have no problems hashing on a single VLAN tagged
> frame.
> > >  However the VLAN will not be a part of the RSS hash.  The  only
> components
> > > of the hash are the IPv4/IPv6 source and destination addresses, and
> if the
> > > flow is TCP then the port numbers.
> > >
> > hi alexander
> > I got these information from the intel community:
> >
> > 'I asked our software engineers about your question, and this is what
> I learned.
> > You cannot filter by just VLAN or VLAN priority.  The L4 type will
> > also play a role in the filter and as such you would only be able to
> > filter TCP, UDP, and SCTP packets that are bound for a VLAN.
> > The command itself to setup a filter is “ethtool –U ethX flow-type
> > tcp4 vlan 0x2000 vlan-mask 0xE000 action Y” where X is the correct
> > index for the interface and Y is the queue you want to route the
> > traffic to.  This would have to be repeated for udp4 and sctp4.
> > I hope this will help.
> > Mark H"
> 
> The mask specifies bits to be ignored, so if you want to filter on the
> basis of only the priority bits you should use vlan-mask 0xfff.  Unless
> this is another inconsistency I failed to notice...
> 
> > so my question is that the VLAN is PART of the RSS or not?
> 
> It's not part of any specified Toeplitz hash.  However, some hardware
> supports adding the hash (after indirection) to the queue number
> specified by a filter.  Currently the ethtool API doesn't have a way to
> request that.
> 
> > looks the
> > perfect filter support vlan id ?can the perfect filter support
> > wildchar,such as: flow-type ANY?
> 
> It is possible to specify this using flow-type ether, but the ixgbe
> driver does not yet support that (and I have no idea whether the
> hardware does).
> 


ethtool –U ethX flow-type tcp4 vlan 0x2000 vlan-mask 0xE000 action Y”

I got this msg:
Cannot add new RX n-tuple filter: Operation not supported
This command is only supported after 2.6.24?

Quoted:

"ethtool: Introduce n-tuple filter programming support

author Peter P Waskiewicz Jr <peter.p.waskiewicz.jr@intel.com>  
 Thu, 11 Feb 2010 04:03:05 +0000 (20:03 -0800) 
committer David S. Miller <davem@davemloft.net>  
 Thu, 11 Feb 2010 04:03:05 +0000 (20:03 -0800) 

This patchset enables the ethtool layer to program n-tuple
filters to an underlying device.  The idea is to allow capable
hardware to have static rules applied that can assist steering
flows into appropriate queues.

Hardware that is known to support these types of filters today
are ixgbe and niu.

Signed-off-by: Peter P Waskiewicz Jr <peter.p.waskiewicz.jr@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
"


> Ben.
> 
> --
> Ben Hutchings, Senior Software Engineer, Solarflare Communications
> Not speaking for my employer; that's the marketing department's job.
> They asked us to note that Solarflare product names are trademarked.


^ permalink raw reply	[flat|nested] 8+ messages in thread

* Re: does intel X520-SR(ixgbe) support RSS on single VLAN?
  2011-01-25  3:05     ` Ben Hutchings
  2011-01-25  7:22       ` Jon Zhou
@ 2011-01-25  9:03       ` Rui
  2011-01-25 21:34         ` Alexander Duyck
  1 sibling, 1 reply; 8+ messages in thread
From: Rui @ 2011-01-25  9:03 UTC (permalink / raw)
  To: Ben Hutchings; +Cc: netdev, e1000-devel

On Tue, Jan 25, 2011 at 11:05 AM, Ben Hutchings
<bhutchings@solarflare.com> wrote:
> On Tue, 2011-01-25 at 10:10 +0800, Rui wrote:
>> On Tue, Jan 25, 2011 at 1:09 AM, Alexander Duyck
>> <alexander.h.duyck@intel.com> wrote:
>> > On 1/24/2011 6:18 AM, Rui wrote:
>> >>
>> >> hi
>> >> does intel X520-SR support RSS on single VLAN?
>> >>
>> >> tested with 3 different vlan id and priority packets
>> >> What I saw is that all packets were always delivered to the same RxQ.
>> >> looks can not get a different RSS index for these packet?
>> >> any setting needed?
>> >> --
>> >> To unsubscribe from this list: send the line "unsubscribe netdev" in
>> >> the body of a message to majordomo@vger.kernel.org
>> >> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>> >
>> > The X520 should have no problems hashing on a single VLAN tagged frame.
>> >  However the VLAN will not be a part of the RSS hash.  The  only components
>> > of the hash are the IPv4/IPv6 source and destination addresses, and if the
>> > flow is TCP then the port numbers.
>> >
>> hi alexander
>> I got these information from the intel community:
>>
>> 'I asked our software engineers about your question, and this is what I learned.
>> You cannot filter by just VLAN or VLAN priority.  The L4 type will
>> also play a role in the filter and as such you would only be able to
>> filter TCP, UDP, and SCTP packets that are bound for a VLAN.
>> The command itself to setup a filter is “ethtool –U ethX flow-type
>> tcp4 vlan 0x2000 vlan-mask 0xE000 action Y” where X is the correct
>> index for the interface and Y is the queue you want to route the
>> traffic to.  This would have to be repeated for udp4 and sctp4.
>> I hope this will help.
>> Mark H"
>
> The mask specifies bits to be ignored, so if you want to filter on the
> basis of only the priority bits you should use vlan-mask 0xfff.  Unless
> this is another inconsistency I failed to notice...
>
>> so my question is that the VLAN is PART of the RSS or not?
>
> It's not part of any specified Toeplitz hash.  However, some hardware
> supports adding the hash (after indirection) to the queue number
> specified by a filter.  Currently the ethtool API doesn't have a way to
> request that.
>
>> looks the
>> perfect filter support vlan id ?can the perfect filter support
>> wildchar,such as: flow-type ANY?
>
> It is possible to specify this using flow-type ether, but the ixgbe
> driver does not yet support that (and I have no idea whether the
> hardware does).
>
> Ben.
>
> --
> Ben Hutchings, Senior Software Engineer, Solarflare Communications
> Not speaking for my employer; that's the marketing department's job.
> They asked us to note that Solarflare product names are trademarked.
>
>

I got this msg:
Cannot add new RX n-tuple filter: Operation not supported
This command is only supported after 2.6.34?

------------------------------------------------------------------------------
Special Offer-- Download ArcSight Logger for FREE (a $49 USD value)!
Finally, a world-class log management solution at an even better price-free!
Download using promo code Free_Logger_4_Dev2Dev. Offer expires 
February 28th, so secure your free ArcSight Logger TODAY! 
http://p.sf.net/sfu/arcsight-sfd2d
_______________________________________________
E1000-devel mailing list
E1000-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/e1000-devel
To learn more about Intel&#174; Ethernet, visit http://communities.intel.com/community/wired

^ permalink raw reply	[flat|nested] 8+ messages in thread

* Re: does intel X520-SR(ixgbe) support RSS on single VLAN?
  2011-01-25  9:03       ` Rui
@ 2011-01-25 21:34         ` Alexander Duyck
  0 siblings, 0 replies; 8+ messages in thread
From: Alexander Duyck @ 2011-01-25 21:34 UTC (permalink / raw)
  To: Rui; +Cc: Ben Hutchings, netdev, e1000-devel

On 1/25/2011 1:03 AM, Rui wrote:
> On Tue, Jan 25, 2011 at 11:05 AM, Ben Hutchings
> <bhutchings@solarflare.com>  wrote:
>> On Tue, 2011-01-25 at 10:10 +0800, Rui wrote:
>>> On Tue, Jan 25, 2011 at 1:09 AM, Alexander Duyck
>>> <alexander.h.duyck@intel.com>  wrote:
>>>> On 1/24/2011 6:18 AM, Rui wrote:
>>>>>
>>>>> hi
>>>>> does intel X520-SR support RSS on single VLAN?
>>>>>
>>>>> tested with 3 different vlan id and priority packets
>>>>> What I saw is that all packets were always delivered to the same RxQ.
>>>>> looks can not get a different RSS index for these packet?
>>>>> any setting needed?
>>>>> --
>>>>> To unsubscribe from this list: send the line "unsubscribe netdev" in
>>>>> the body of a message to majordomo@vger.kernel.org
>>>>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>>>>
>>>> The X520 should have no problems hashing on a single VLAN tagged frame.
>>>>   However the VLAN will not be a part of the RSS hash.  The  only components
>>>> of the hash are the IPv4/IPv6 source and destination addresses, and if the
>>>> flow is TCP then the port numbers.
>>>>
>>> hi alexander
>>> I got these information from the intel community:
>>>
>>> 'I asked our software engineers about your question, and this is what I learned.
>>> You cannot filter by just VLAN or VLAN priority.  The L4 type will
>>> also play a role in the filter and as such you would only be able to
>>> filter TCP, UDP, and SCTP packets that are bound for a VLAN.
>>> The command itself to setup a filter is “ethtool –U ethX flow-type
>>> tcp4 vlan 0x2000 vlan-mask 0xE000 action Y” where X is the correct
>>> index for the interface and Y is the queue you want to route the
>>> traffic to.  This would have to be repeated for udp4 and sctp4.
>>> I hope this will help.
>>> Mark H"
>>
>> The mask specifies bits to be ignored, so if you want to filter on the
>> basis of only the priority bits you should use vlan-mask 0xfff.  Unless
>> this is another inconsistency I failed to notice...
>>
>>> so my question is that the VLAN is PART of the RSS or not?
>>
>> It's not part of any specified Toeplitz hash.  However, some hardware
>> supports adding the hash (after indirection) to the queue number
>> specified by a filter.  Currently the ethtool API doesn't have a way to
>> request that.
>>
>>> looks the
>>> perfect filter support vlan id ?can the perfect filter support
>>> wildchar,such as: flow-type ANY?
>>
>> It is possible to specify this using flow-type ether, but the ixgbe
>> driver does not yet support that (and I have no idea whether the
>> hardware does).
>>
>> Ben.
>>
>> --
>> Ben Hutchings, Senior Software Engineer, Solarflare Communications
>> Not speaking for my employer; that's the marketing department's job.
>> They asked us to note that Solarflare product names are trademarked.
>>
>>
>
> I got this msg:
> Cannot add new RX n-tuple filter: Operation not supported
> This command is only supported after 2.6.34?

Yes, this command will only work on 2.6.34 and older kernels.

Thanks,

Alex

^ permalink raw reply	[flat|nested] 8+ messages in thread

end of thread, other threads:[~2011-01-25 21:34 UTC | newest]

Thread overview: 8+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2011-01-24 14:18 does intel X520-SR(ixgbe) support RSS on single VLAN? Rui
2011-01-24 17:09 ` Alexander Duyck
2011-01-24 18:10   ` Rick Jones
2011-01-25  2:10   ` Rui
2011-01-25  3:05     ` Ben Hutchings
2011-01-25  7:22       ` Jon Zhou
2011-01-25  9:03       ` Rui
2011-01-25 21:34         ` Alexander Duyck

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.