[dpdk-dev] Duplicating traffic with RTE Flow

Jan Viktorin viktorin at cesnet.cz
Mon Mar 15 14:22:16 CET 2021


Hello Jiawei,

On Fri, 12 Mar 2021 09:32:44 +0000
"Jiawei(Jonny) Wang" <jiaweiw at nvidia.com> wrote:

> Hi Jan,
> 
> > -----Original Message-----
> > From: Jan Viktorin <viktorin at cesnet.cz>
> > Sent: Friday, March 12, 2021 12:33 AM
> > To: Jiawei(Jonny) Wang <jiaweiw at nvidia.com>
> > Cc: Slava Ovsiienko <viacheslavo at nvidia.com>; Asaf Penso
> > <asafp at nvidia.com>; dev at dpdk.org; Ori Kam <orika at nvidia.com>
> > Subject: Re: [dpdk-dev] Duplicating traffic with RTE Flow
> > 
> > On Thu, 11 Mar 2021 02:11:07 +0000
> > "Jiawei(Jonny) Wang" <jiaweiw at nvidia.com> wrote:
> >   
> > > Hi Jan,
> > >
> > > Sorry for late response,
> > >
> > > First rule is invalid, port only works on FDB domain so need
> > > 'transfer' here; Second rule should be ok,  could you please check if the  
> > port 1 was enabled on you dpdk application?
> > 
> > I assume that it is enabled, see full transcript:
> > 
> >  $ ofed_info
> >  MLNX_OFED_LINUX-5.2-1.0.4.0 (OFED-5.2-1.0.4):
> >  ...
> >  $ sudo dpdk-testpmd -v -- -i
> >  EAL: Detected 24 lcore(s)
> >  EAL: Detected 1 NUMA nodes
> >  EAL: RTE Version: 'DPDK 20.11.0'
> >  EAL: Multi-process socket /var/run/dpdk/rte/mp_socket
> >  EAL: Selected IOVA mode 'PA'
> >  EAL: No available hugepages reported in hugepages-1048576kB
> >  EAL: Probing VFIO support...
> >  EAL: Probe PCI driver: mlx5_pci (15b3:1017) device: 0000:04:00.0 (socket 0)
> >  mlx5_pci: No available register for Sampler.
> >  mlx5_pci: Size 0xFFFF is not power of 2, will be aligned to 0x10000.
> >  EAL: Probe PCI driver: mlx5_pci (15b3:1017) device: 0000:04:00.1 (socket 0)
> >  mlx5_pci: No available register for Sampler.
> >  mlx5_pci: Size 0xFFFF is not power of 2, will be aligned to 0x10000.
> >  EAL: No legacy callbacks, legacy socket not created  Interactive-mode
> > selected
> >  testpmd: create a new mbuf pool <mb_pool_0>: n=331456, size=2176,
> > socket=0
> >  testpmd: preferred mempool ops selected: ring_mp_mc  Configuring Port 0
> > (socket 0)  Port 0: B8:59:9F:E2:09:F6  Configuring Port 1 (socket 0)  Port 1:
> > B8:59:9F:E2:09:F7  Checking link statuses...
> >  Done  
> 
> Seems that you start two PF port here,  Port 1 is not VF port;
> FDB rule can steering the packet form PF to its VFs and vice versa, Could you please try to open the
> VF ports and start the testpmd with representor=.

I did not know this, so I tried with VFs:

 # echo 2 > /sys/class/net/hge1/device/sriov_numvfs
 # echo switchdev > /sys/class/net/hge1/compat/devlink/mode

 # dpdk-testpmd -v -a '0000:05:00.1,representor=[0-1]' -- -i
 EAL: Detected 24 lcore(s)
 EAL: Detected 1 NUMA nodes
 EAL: RTE Version: 'DPDK 20.11.0'
 EAL: Multi-process socket /var/run/dpdk/rte/mp_socket
 EAL: Selected IOVA mode 'VA'
 EAL: No available hugepages reported in hugepages-1048576kB
 EAL: Probing VFIO support...
 EAL: Probe PCI driver: mlx5_pci (15b3:1017) device: 0000:05:00.1 (socket 0)
 mlx5_pci: No available register for Sampler.
 mlx5_pci: Size 0xFFFF is not power of 2, will be aligned to 0x10000.
 mlx5_pci: No available register for Sampler.
 mlx5_pci: No available register for Sampler.
 EAL: No legacy callbacks, legacy socket not created
 Interactive-mode selected
 testpmd: create a new mbuf pool <mb_pool_0>: n=331456, size=2176, socket=0
 testpmd: preferred mempool ops selected: ring_mp_mc

 Warning! port-topology=paired and odd forward ports number, the last port will pair with itself.

 Configuring Port 0 (socket 0)
 Port 0: B8:59:9F:E2:09:F7
 Configuring Port 1 (socket 0)
 Port 1: B2:57:D6:72:F3:31
 Configuring Port 2 (socket 0)
 Port 2: 9E:CB:D0:73:59:CE
 Checking link statuses...
 Done
 testpmd> show port summary all
 Number of available ports: 3
 Port MAC Address       Name         Driver         Status   Link
 0    B8:59:9F:E2:09:F7 0000:05:00.1 mlx5_pci       up       100 Gbps
 1    B2:57:D6:72:F3:31 0000:05:00.1_representor_0 mlx5_pci       up       100 Gbps
 2    9E:CB:D0:73:59:CE 0000:05:00.1_representor_1 mlx5_pci       up       100 Gbps
 testpmd> set sample_actions 0 port_id id 1 / end
 testpmd> flow validate 0 ingress transfer pattern end actions sample ratio 1 index 0 / drop / end
 port_flow_complain(): Caught PMD error type 1 (cause unspecified): sample action not supported: Operation not supported

Still no luck. However, there is this message 3-times in the log:

 mlx5_pci: No available register for Sampler.

It looks like it might be related. What does it mean?

Jan

> 
> Thanks.
> 
> >  testpmd> port start 1  
> >  Port 1 is now not stopped
> >  Please stop the ports first
> >  Done  
> >  testpmd> set sample_actions 0 port_id id 1 / end  testpmd> flow validate 0  
> > ingress transfer pattern end actions sample ratio 1 index 0 / drop / end
> >  port_flow_complain(): Caught PMD error type 1 (cause unspecified): (no
> > stated reason): Operation not supported  testpmd> flow create 0 ingress
> > transfer pattern end actions sample ratio 1 index 0 / drop / end
> >  port_flow_complain(): Caught PMD error type 1 (cause unspecified): (no
> > stated reason): Operation not supported  testpmd>  Stopping port 0...
> >  Stopping ports...
> >  Done
> > 
> >  Stopping port 1...
> >  Stopping ports...
> >  Done
> > 
> >  Shutting down port 0...
> >  Closing ports...
> >  Port 0 is closed
> >  Done
> > 
> >  Shutting down port 1...
> >  Closing ports...
> >  Port 1 is closed
> >  Done
> > 
> >  Bye...
> > 
> > Jan
> >   
> > >
> > > Thanks.
> > > Jonny
> > >  
> > > > -----Original Message-----
> > > > From: Jan Viktorin <viktorin at cesnet.cz>
> > > > Sent: Monday, March 1, 2021 10:43 PM
> > > > To: Slava Ovsiienko <viacheslavo at nvidia.com>
> > > > Cc: Asaf Penso <asafp at nvidia.com>; dev at dpdk.org; Ori Kam
> > > > <orika at nvidia.com>; Jiawei(Jonny) Wang <jiaweiw at nvidia.com>
> > > > Subject: Re: [dpdk-dev] Duplicating traffic with RTE Flow
> > > >
> > > > On Mon, 1 Mar 2021 14:34:07 +0000
> > > > Slava Ovsiienko <viacheslavo at nvidia.com> wrote:
> > > >  
> > > > > Hi, Jan
> > > > >
> > > > > To use port action (I see it is in your sample action list) the
> > > > > flow should be applied to the FDB domain, ie "transfer" attribute
> > > > > should be  
> > > > specified:  
> > > > >
> > > > > flow validate 0 ingress transfer...  
> > > >
> > > > As you can see (however, it's a bit messy in the response below, in
> > > > [1], it is better), I tried both. First without transfer and second
> > > > with. The first gives hint "action is valid in transfer mode only"
> > > > but the second try with transfer gives "Operation not supported".
> > > >
> > > > Jan
> > > >
> > > > [1] http://mails.dpdk.org/archives/dev/2021-March/200475.html
> > > >  
> > > > >
> > > > > With best regards, Slava
> > > > >  
> > > > > > -----Original Message-----
> > > > > > From: Jan Viktorin <viktorin at cesnet.cz>
> > > > > > Sent: Monday, March 1, 2021 14:21
> > > > > > To: Asaf Penso <asafp at nvidia.com>
> > > > > > Cc: dev at dpdk.org; Ori Kam <orika at nvidia.com>; Jiawei(Jonny) Wang
> > > > > > <jiaweiw at nvidia.com>; Slava Ovsiienko <viacheslavo at nvidia.com>
> > > > > > Subject: Re: [dpdk-dev] Duplicating traffic with RTE Flow
> > > > > >
> > > > > > Hello Asaf,
> > > > > >
> > > > > > it is a while we were in touch regarding this topic. Finally, I
> > > > > > am again trying to get work this feature. I've seen that
> > > > > > sampling is already upstreamed which is great. However, I am not
> > > > > > very successful with that. There is nearly no documentation,
> > > > > > just [1], I found no examples,  
> > > > just commit logs...  
> > > > > >
> > > > > > I tried:
> > > > > >  
> > > > > >  > set sample_actions 0 port_id id 1 / end  > flow validate 0  
> > > > > > ingress pattern end actions sample ratio 1 index 0 / drop / end
> > > > > >  port_flow_complain(): Caught PMD error type 1 (cause unspecified):
> > > > > > port id action is valid in transfer mode only: Operation not
> > > > > > supported  > flow validate
> > > > > > 0 ingress transfer pattern end actions sample ratio 1 index 0 /
> > > > > > drop / end
> > > > > >  port_flow_complain(): Caught PMD error type 1 (cause unspecified):
> > > > > > (no stated reason): Operation not supported
> > > > > >
> > > > > > Using CentOS 7, DPDK 20.11.0, OFED-5.2-1.0.4.
> > > > > > NICs: MT2892 Family [ConnectX-6 Dx] 101d (fw 22.28.1002),
> > > > > > MT27800 Family [ConnectX-5] 1017 (fw 16.27.2008).
> > > > > >
> > > > > > My primary goal is to be able to deliver exactly the same
> > > > > > packets both to DPDK and to the Linux kernel. Doing this at RTE
> > > > > > Flow level would be great due to performance and transparency.
> > > > > >
> > > > > > Jan
> > > > > >
> > > > > > [1]
> > > > > > https://doc.dpdk.org/guides/prog_guide/rte_flow.html#action-samp
> > > > > > le
> > > > > >
> > > > > > On Fri, 18 Sep 2020 14:23:42 +0000 Asaf Penso <asafp at nvidia.com>
> > > > > > wrote:
> > > > > >  
> > > > > > > Hello Jan,
> > > > > > >
> > > > > > > You can have a look in series [1] where we propose to add APIs
> > > > > > > to  
> > > > > > DPDK20.11 for both mirroring and sampling for packets, with
> > > > > > additional actions of the different traffic.  
> > > > > > >
> > > > > > > [1]
> > > > > > > http://patches.dpdk.org/project/dpdk/list/?series=12045
> > > > > > >
> > > > > > > Regards,
> > > > > > > Asaf Penso
> > > > > > >  
> > > > > > > >-----Original Message-----
> > > > > > > >From: dev <dev-bounces at dpdk.org> On Behalf Of Jan Viktorin
> > > > > > > >Sent: Friday, September 18, 2020 3:56 PM
> > > > > > > >To: dev at dpdk.org
> > > > > > > >Subject: [dpdk-dev] Duplicating traffic with RTE Flow
> > > > > > > >
> > > > > > > >Hello all,
> > > > > > > >
> > > > > > > >we are looking for a way to duplicate ingress traffic in hardware.
> > > > > > > >
> > > > > > > >There is an example in [1] suggesting to insert two fate
> > > > > > > >actions into the RTE Flow actions array like:
> > > > > > > >
> > > > > > > >  flow create 0 ingress pattern end \
> > > > > > > >      actions queue index 0 / void / queue index 1 / end
> > > > > > > >
> > > > > > > >But our experience is that PMDs reject two fate actions
> > > > > > > >(tried with mlx5). Another similar approach would be to
> > > > > > > >deliver every single packet into two virtual
> > > > > > > >functions:
> > > > > > > >
> > > > > > > >  flow create 0 ingress pattern end \
> > > > > > > >     actions vf index 0 / vf index 1 / end
> > > > > > > >
> > > > > > > >Third possibility was to use passthru:
> > > > > > > >
> > > > > > > >  flow create 0 ingress pattern end \
> > > > > > > >      actions passthru / vf index 0 / end  flow create 0
> > > > > > > > ingress pattern end \
> > > > > > > >      actions vf index 1 / end
> > > > > > > >
> > > > > > > >Again, tried on mlx5 and it does not support the passthru.
> > > > > > > >
> > > > > > > >Last idea was to use isolate with passthru (to deliver both
> > > > > > > >to DPDK application and to the kernel) but again there was no
> > > > > > > >support on mlx5 for  
> > > > > > passthru...  
> > > > > > > >
> > > > > > > >  flow isolate 0 true
> > > > > > > >  flow create 0 ingress pattern end actions passthru / rss
> > > > > > > > end / end
> > > > > > > >
> > > > > > > >Is there any other possibility or PMD+NIC that is known to
> > > > > > > >solve such  
> > > > > > issue?  
> > > > > > > >
> > > > > > > >Thanks
> > > > > > > >Jan Viktorin
> > > > > > > >
> > > > > > > >[1]
> > > > > > > >https://nam11.safelinks.protection.outlook.com/?url=https%3A%
> > > > > > > >2F%  
> > > > 2  
> > > > > > > >Fdoc
> > > > > > > >.dpdk
> > > > > > > >.org%2Fguides%2Fprog_guide%2Frte_flow.html%23table-rte-flow-  
> > > > redir  
> > > > > > > >ect-
> > > > > > > >queue-5-  
> > > > > > >
> > > > > > >3&data=02%7C01%7Casafp%40nvidia.com%7C1a46005bec5245  
> > e72  
> > > > 9e70  
> > > > > > 8d  
> > > > > > >
> > > > > > >85bd24caf%7C43083d15727340c1b7db39efd9ccc17a%7C0%7C0%7C6  
> > 373  
> > > > 6030  
> > > > > > 60  
> > > > > > >
> > > > > > >73519816&sdata=EOF%2Fz62crvBZK8rwzwKIWxj5cVlfPVnU3FLm  
> > cL9  
> > > > X2w0  
> > > > > > %3  
> > > > > > > >D&reserved=0  
> > > > >  
> > >  
> 



More information about the dev mailing list