Re: [vpp-dev] is VPP Gerrit server "No space left on device"

[Xu, Ting]

Hi,

+1, Meet the same issue

From: vpp-dev@lists.fd.io  On Behalf Of Pei, Yulong
Sent: Monday, January 30, 2023 10:12 AM
To: vpp-dev@lists.fd.io
Subject: Re: [vpp-dev] is VPP Gerrit server "No space left on device"
Importance: High

When open patch link of gerrit, it pops up below message with web browser, and 
can not find patch contents any more,

[cid:image001.png@01D93495.88F465F0]

From: vpp-dev@lists.fd.io 
mailto:vpp-dev@lists.fd.io>> On Behalf Of Pei, Yulong
Sent: Saturday, January 28, 2023 10:30 PM
To: vpp-dev mailto:vpp-dev@lists.fd.io>>
Subject: [vpp-dev] is VPP Gerrit server "No space left on device"

Hello vpp-dev,

Failed to run `git review`,  is VPP Gerrit server "No space left on device" ?

# git review
error: remote unpack failed: error No space left on device
fatal: Unpack error, check server log
To ssh://gerrit.fd.io:29418/vpp

-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#22515): https://lists.fd.io/g/vpp-dev/message/22515
Mute This Topic: https://lists.fd.io/mt/96587880/21656
Group Owner: vpp-dev+ow...@lists.fd.io
Unsubscribe: https://lists.fd.io/g/vpp-dev/leave/1480452/21656/631435203/xyzzy 
[arch...@mail-archive.com]
-=-=-=-=-=-=-=-=-=-=-=-

Re: [vpp-dev] Review request for Generic Flow in Native AVF

[Xu, Ting]

CSIT check issue is resolved.

From: Xu, Ting
Sent: Tuesday, November 29, 2022 8:23 PM
To: vpp-dev 
Cc: Damjan Marion 
Subject: Review request for Generic Flow in Native AVF

Hi, Damjan and community,

I have uploaded a patch for enabling generic flow in native AVF days ago.
To support generic flow in native AVF, I changed the original virtual channel 
message handling method of FDIR, and add RSS hash support as well. Since Damjan 
you are the maintainer of native AVF module, I really need your comments on 
whether there is any problems of my modification.

There is one key issue I noticed recently, that is I just enabled RSS hash for 
generic flow, but not for the common flow type. Do you think it is OK to do so? 
And it may cause a problem of CSIT test for flow in native AVF. Does anyone 
have any advice?

Thanks!

The link for the patch is 37563: avf: support generic flow | 
https://gerrit.fd.io/r/c/vpp/+/37563

Best Regards,
Xu Ting

-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#22261): https://lists.fd.io/g/vpp-dev/message/22261
Mute This Topic: https://lists.fd.io/mt/95332823/21656
Group Owner: vpp-dev+ow...@lists.fd.io
Unsubscribe: https://lists.fd.io/g/vpp-dev/leave/1480452/21656/631435203/xyzzy 
[arch...@mail-archive.com]
-=-=-=-=-=-=-=-=-=-=-=-

[vpp-dev] Review request for Generic Flow in Native AVF

[Xu, Ting]

Hi, Damjan and community,

I have uploaded a patch for enabling generic flow in native AVF days ago.
To support generic flow in native AVF, I changed the original virtual channel 
message handling method of FDIR, and add RSS hash support as well. Since Damjan 
you are the maintainer of native AVF module, I really need your comments on 
whether there is any problems of my modification.

There is one key issue I noticed recently, that is I just enabled RSS hash for 
generic flow, but not for the common flow type. Do you think it is OK to do so? 
And it may cause a problem of CSIT test for flow in native AVF. Does anyone 
have any advice?

Thanks!

The link for the patch is 37563: avf: support generic flow | 
https://gerrit.fd.io/r/c/vpp/+/37563

Best Regards,
Xu Ting

-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#22254): https://lists.fd.io/g/vpp-dev/message/22254
Mute This Topic: https://lists.fd.io/mt/95332823/21656
Group Owner: vpp-dev+ow...@lists.fd.io
Unsubscribe: https://lists.fd.io/g/vpp-dev/leave/1480452/21656/631435203/xyzzy 
[arch...@mail-archive.com]
-=-=-=-=-=-=-=-=-=-=-=-

Re: [vpp-dev] RFC: move the “1”st instance of the community meeting 3 hours earlier to reduce impact for earlier time zone

[Xu, Ting]

+1, it would be convenient for us to attend the meeting!

From: vpp-dev@lists.fd.io  On Behalf Of Pei, Yulong
Sent: Thursday, November 24, 2022 10:19 AM
To: vpp-dev@lists.fd.io
Cc: Li, Jokul 
Subject: Re: [vpp-dev] RFC: move the “1”st instance of the community meeting 3 
hours earlier to reduce impact for earlier time zone

+1, Looking forward to join community meeting.

From: vpp-dev@lists.fd.io 
mailto:vpp-dev@lists.fd.io>> On Behalf Of Ni, Hongjun
Sent: Wednesday, November 23, 2022 11:51 AM
To: vpp-dev@lists.fd.io
Cc: Li, Jokul mailto:jokul...@intel.com>>
Subject: Re: [vpp-dev] RFC: move the “1”st instance of the community meeting 3 
hours earlier to reduce impact for earlier time zone

+1. Hope more PRC guys will join and contribute to VPP community!

From: vpp-dev@lists.fd.io 
mailto:vpp-dev@lists.fd.io>> On Behalf Of qian xu
Sent: Wednesday, November 23, 2022 9:28 AM
To: vpp-dev@lists.fd.io
Cc: Li, Jokul mailto:jokul...@intel.com>>
Subject: Re: [vpp-dev] RFC: move the “1”st instance of the community meeting 3 
hours earlier to reduce impact for earlier time zone

+1, thanks for the inclusion for PRC contributors, looking forwards to joining 
the community meetings in near future!

From: vpp-dev@lists.fd.io 
mailto:vpp-dev@lists.fd.io>> On Behalf Of Dave Wallace
Sent: Wednesday, November 23, 2022 2:48 AM
To: vpp-dev@lists.fd.io
Subject: Re: [vpp-dev] RFC: move the “1”st instance of the community meeting 3 
hours earlier to reduce impact for earlier time zone

+1
On 11/22/22 11:28 AM, Andrew Yourtchenko wrote:

Hi all,



It came up that the current time of the community meeting is extremely 
unfriendly for our community members on China timezone - it makes it midnight 
their time.



So, as per discussion on the call today, I would like to propose to move every 
other of the meetings (the one in the second Tuesday of the month) three hours 
earlier.



I would also suggest that this proposal take effect starting with the upcoming 
meeting on December 13th.



Thoughts ?



--a









-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#22239): https://lists.fd.io/g/vpp-dev/message/22239
Mute This Topic: https://lists.fd.io/mt/9526/21656
Group Owner: vpp-dev+ow...@lists.fd.io
Unsubscribe: https://lists.fd.io/g/vpp-dev/leave/1480452/21656/631435203/xyzzy 
[arch...@mail-archive.com]
-=-=-=-=-=-=-=-=-=-=-=-

[vpp-dev] RFC for generic flow in native avf

[Xu, Ting]

Hi,

I submitted a patch for supporting generic flow in native avf.  
https://gerrit.fd.io/r/c/vpp/+/37563

Generic flow related features were discussed and accepted in the past two 
releases, for example, generic flow in dpdk plugins and packetforge. This time 
we are willing to enable it in native avf.
However, since native avf does not support RSS now, we should also enable RSS 
function this time. It is the reason why we mark this patch as RFC.

We only enabled RSS functions needed by generic flow in this patch. It will 
parse the generic flow patterns and deliver the spec and mask to kernel driver 
through virtual channel. The same in FDIR. Native avf does not need to 
translate generic flow to hardware accepted flow profile. It is the work of 
kernel driver. We kept a position for common RSS process, but it will do 
nothing.

Since the original "avf_flow_add" functions are only for FDIR, we extend them 
for RSS and FDIR both. The flow rule type (RSS or FDIR) will be determined by 
the flow command, for example, the "test flow" command in vnet/flow CLI, 
"redirect-to-queue 3" means FDIR and "rss function default" means RSS.

[cid:image002.png@01D8EDDF.A64C5560]

I would be grateful if maintainers can have a look at this RFC patch, and give 
comments especially on RSS design in native avf. Thanks!

Best Regards
Xu Ting


image001.emz
Description: image001.emz


oledata.mso
Description: oledata.mso

-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#22110): https://lists.fd.io/g/vpp-dev/message/22110
Mute This Topic: https://lists.fd.io/mt/94701971/21656
Group Owner: vpp-dev+ow...@lists.fd.io
Unsubscribe: https://lists.fd.io/g/vpp-dev/leave/1480452/21656/631435203/xyzzy 
[arch...@mail-archive.com]
-=-=-=-=-=-=-=-=-=-=-=-

Re: [vpp-dev] Reminder: 7 days to RC1 milestone of VPP 22.10

[Xu, Ting]

Hi, Andrew

The latest version is the one modified according to the comments.
I deleted most of the samples, left a few typical ones, and write the guidance 
how to create the json profile in README and Spec

Thanks

-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#21893): https://lists.fd.io/g/vpp-dev/message/21893
Mute This Topic: https://lists.fd.io/mt/93678303/21656
Group Owner: vpp-dev+ow...@lists.fd.io
Unsubscribe: https://lists.fd.io/g/vpp-dev/leave/1480452/21656/631435203/xyzzy 
[arch...@mail-archive.com]
-=-=-=-=-=-=-=-=-=-=-=-

Re: [vpp-dev] Reminder: 7 days to RC1 milestone of VPP 22.10

[Xu, Ting]

Hi, Andrew

My patch https://gerrit.fd.io/r/c/vpp/+/36149 (packetforge) is waiting to be 
merged, thanks.

Best Regards,
Xu Ting

-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#21882): https://lists.fd.io/g/vpp-dev/message/21882
Mute This Topic: https://lists.fd.io/mt/93678303/21656
Group Owner: vpp-dev+ow...@lists.fd.io
Unsubscribe: https://lists.fd.io/g/vpp-dev/leave/1480452/21656/631435203/xyzzy 
[arch...@mail-archive.com]
-=-=-=-=-=-=-=-=-=-=-=-

[vpp-dev] One proposal for community meeting

[Xu, Ting]

Hi,

As I see in Dave's message, our Community Meeting will be held next Tuesday 
(9.13).

I am wondering if I can have some time to present my patch - packetforge (5 
minutes). It's been a long time since it was last reviewed, and I think perhaps 
it is not clear to understand.

The patch link is: https://gerrit.fd.io/r/c/vpp/+/36149

BTW, should I follow the step here https://wiki.fd.io/view/VPP/Meeting to join 
the meeting?

Thanks!

Best Regards,
Xu Ting

-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#21848): https://lists.fd.io/g/vpp-dev/message/21848
Mute This Topic: https://lists.fd.io/mt/93494249/21656
Group Owner: vpp-dev+ow...@lists.fd.io
Unsubscribe: https://lists.fd.io/g/vpp-dev/leave/1480452/21656/631435203/xyzzy 
[arch...@mail-archive.com]
-=-=-=-=-=-=-=-=-=-=-=-

Re: [vpp-dev] Gerrit review request for packetforge - 36149

[Xu, Ting]

I have refined the packetforge script and add the function to disable/delete 
flow rules. Please take a look and review if convenient, thanks!

https://gerrit.fd.io/r/c/vpp/+/36149

-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#21818): https://lists.fd.io/g/vpp-dev/message/21818
Mute This Topic: https://lists.fd.io/mt/92596268/21656
Group Owner: vpp-dev+ow...@lists.fd.io
Unsubscribe: https://lists.fd.io/g/vpp-dev/leave/1480452/21656/631435203/xyzzy 
[arch...@mail-archive.com]
-=-=-=-=-=-=-=-=-=-=-=-

Re: [vpp-dev] Gerrit review request for packetforge - 36149

[Xu, Ting]

Hi, Pim

We have a readme file in the patch, and I attached it here.

BTW, the corresponding patch for generic flow is 
https://gerrit.fd.io/r/c/vpp/+/34802

Best Regards,
Xu Ting

From: vpp-dev@lists.fd.io  On Behalf Of Pim van Pelt
Sent: Monday, July 25, 2022 9:09 AM
To: vpp-dev@lists.fd.io
Cc: Damjan Marion 
Subject: Re: [vpp-dev] Gerrit review request for packetforge - 36149

Hi Xu,

Do you have documentation (notably: a userguide) for this tool? The readme is 
sparse and it seems like "read the code" is a bit much to ask :)

groet,
Pim

On Sun, Jul 24, 2022 at 6:04 PM Xu, Ting 
mailto:ting...@intel.com>> wrote:
Hi,

I have submitted a patch for packetforge weeks ago 
(https://gerrit.fd.io/r/c/vpp/+/36149). It is a tool to create flow rules using 
generic flow, which was upstreamed in the last release.

Packetforge can provide a user-friendly interface to add/enable flow rules to 
use NIC’s hardware flow offloading function.

According to Damjan’s comment, we put packetforge in ./extras, and in this v1 
patch, we created a python script as a interface, to support add and enable 
function. We will refine it and add more functions in the future.

The detailed description and usage can be found in commit log and README in the 
patch.

Please have a look at my patch and give some comments, thanks!

Best Regards,
Xu Ting




--
Pim van Pelt mailto:p...@ipng.nl>>
PBVP1-RIPE - http://www.ipng.nl/
.. _packetforge_doc:

Packetforge for generic flow


Packetforge is a tool to support generic flow. Since the input format of
generic flow is hard to read and create, packetforge can help to create
generic flow rules using a format of naming protocols (like Scapy) or json
profile. Packetforge is built based on a parsegraph, users can modify the
graph nodes and edges if needed.

Command examples


::

 $ python flow_create.py -p "mac()/ipv4(src=1.1.1.1,dst=2.2.2.2)/udp()"
   -a "redirect-to-queue 3"

Naming format input. It needs two parameters. Pattern format is similar to
Scapy protocols. Actions format is the same as vnet/flow command. This
command will add and enable flow rules to interface 1.

::

 $ python flow_create.py -f "./flow_rule_examples/mac_ipv4.json"

Json profile format input. This command takes a json profile as parameter.
In the json profile, there will be protocols and their fields and values.
Users can define spec and mask for each field. Actions can be added in the
profile directly, otherwise "-a" option should be added in the command.
The example can be found in flow_rule_examples folder.

::

  $ show flow entry

It is a vnet/flow command, used in VPP CLI. It can show the added flow rules
after using the above commands. In addition, if users want to delete or
disable the rules, vnet/flow commands "test flow del" and "test flow disable"
are valid.

ParseGraph
--

Packetforge is built based on a ParseGraph. The ParseGraph is constructed
with nodes and edges. Nodes are protocols, including information about
protocol's name, fields and default values. Edges are the relationship
between two protocols, including some actions needed when connecting two
protocols. For example, change the mac header ethertype to 0x0800 when
connecting mac and ipv4. More details are in the spec in parsegraph folder.
Users can build the ParseGraph following the spec by themselves, like
adding a new protocol. If NIC supports the new protocol, the rule can be
created. Otherwise, it will return error.

-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#21719): https://lists.fd.io/g/vpp-dev/message/21719
Mute This Topic: https://lists.fd.io/mt/92596268/21656
Group Owner: vpp-dev+ow...@lists.fd.io
Unsubscribe: https://lists.fd.io/g/vpp-dev/leave/1480452/21656/631435203/xyzzy 
[arch...@mail-archive.com]
-=-=-=-=-=-=-=-=-=-=-=-

[vpp-dev] Gerrit review request for packetforge - 36149

[Xu, Ting]

Hi,

I have submitted a patch for packetforge weeks ago 
(https://gerrit.fd.io/r/c/vpp/+/36149). It is a tool to create flow rules using 
generic flow, which was upstreamed in the last release.

Packetforge can provide a user-friendly interface to add/enable flow rules to 
use NIC's hardware flow offloading function.

According to Damjan's comment, we put packetforge in ./extras, and in this v1 
patch, we created a python script as a interface, to support add and enable 
function. We will refine it and add more functions in the future.

The detailed description and usage can be found in commit log and README in the 
patch.

Please have a look at my patch and give some comments, thanks!

Best Regards,
Xu Ting

-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#21716): https://lists.fd.io/g/vpp-dev/message/21716
Mute This Topic: https://lists.fd.io/mt/92596268/21656
Group Owner: vpp-dev+ow...@lists.fd.io
Unsubscribe: https://lists.fd.io/g/vpp-dev/leave/1480452/21656/631435203/xyzzy 
[arch...@mail-archive.com]
-=-=-=-=-=-=-=-=-=-=-=-

[vpp-dev] Support generic flow and RSS action in flow VAPI

[Xu, Ting]

Hi,

I submit a patch in vnet flow to support generic flow and RSS action in VAPI. 
It is a part of work to enable generic flow in vnet flow. It has been there for 
a long time. Could anyone help to review?

Link: https://gerrit.fd.io/r/c/vpp/+/35614

Thanks so much!

Best Regards,
Xu Ting

-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#21394): https://lists.fd.io/g/vpp-dev/message/21394
Mute This Topic: https://lists.fd.io/mt/91027989/21656
Group Owner: vpp-dev+ow...@lists.fd.io
Unsubscribe: https://lists.fd.io/g/vpp-dev/unsub [arch...@mail-archive.com]
-=-=-=-=-=-=-=-=-=-=-=-

Re: [vpp-dev] Segmentation fault when dpdk number-rx-queues > 1 in startup.conf

[Xu, Ting]

Hi, Damjan

I look into the code. The bad commit is 
ce4083ce48958d9d3956e8317445a5552780af1a (“dpdk: offloads cleanup”), and the 
previous commit is correct, so I compare these two. Since they use the same 
DPDK version, I check the input of rte API.

I find that the direct cause is configuring default RSS in DPDK. It is called 
by *dpdk_device_setup()* in dpdk plugins, the API function is 
*rte_eth_dev_configure()*. However, the bad commit and the good commit have 
almost the same input to *rte_eth_dev_configure()* , the only difference is a 
Tx offload flag (TX_IPV4_CSUM), but I think it is not the root cause because it 
does not help after I fix it. Since they have the same input to dpdk API, I 
think it is not DPDK's issue.

I find there are a lot of flags or offloads configuring change in commit 
(“dpdk: offloads cleanup”). I guess is it possible that some flags are not 
correct? I look at the code in *dpdk_lib_init()* , but do not find the cause 
yet.

Do you have any suggestion to me? Thanks!

Best Regards,
Xu Ting

-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#21379): https://lists.fd.io/g/vpp-dev/message/21379
Mute This Topic: https://lists.fd.io/mt/89520993/21656
Group Owner: vpp-dev+ow...@lists.fd.io
Unsubscribe: https://lists.fd.io/g/vpp-dev/unsub [arch...@mail-archive.com]
-=-=-=-=-=-=-=-=-=-=-=-

[vpp-dev] RSS queue group action in vnet/flow CLI

[Xu, Ting]

Hi,

I have a patch sent 2 weeks ago, which is about to enable RSS queue group 
action in vnet/flow CLI, but no body review. The patch is below:
https://gerrit.fd.io/r/c/vpp/+/35780

RSS queue group action is a supported action in DPDK. It can distribute packets 
among several assigned continuous queues. It is missed in vnet/flow when 
creating this flow frame. I add the command parsing process to support flow CLI 
to receive and parse the request for RSS queue group action. The request will 
be passed to DPDK rte_flow. It is also a support for 5G enhancement. Please 
help review and give your advice, thanks!

Best Regards,
Xu Ting


-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#21257): https://lists.fd.io/g/vpp-dev/message/21257
Mute This Topic: https://lists.fd.io/mt/90456279/21656
Group Owner: vpp-dev+ow...@lists.fd.io
Unsubscribe: https://lists.fd.io/g/vpp-dev/unsub [arch...@mail-archive.com]
-=-=-=-=-=-=-=-=-=-=-=-

[vpp-dev] How to add compile flag to VPP makefile

[Xu, Ting]

Hi,

I am just doing a test and need to include Python.h in the vnet/pg code. When I 
"make build", it reports error that Python.h is not found.
In my environment, I find out that I need to add the flag 
"-I/usr/include/Python3.9 -lpython3.9" when I compile a code with Python.h 
using gcc directly. However, my attempt to add it to VPP failed.

Therefore, I would like to know that how to add compilation flags 
"-I/usr/include/Python3.9 -lpython3.9" to the correct compilation file in VPP?

Thanks!

Best Regards,
Xu Ting

-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#21247): https://lists.fd.io/g/vpp-dev/message/21247
Mute This Topic: https://lists.fd.io/mt/90436226/21656
Group Owner: vpp-dev+ow...@lists.fd.io
Unsubscribe: https://lists.fd.io/g/vpp-dev/unsub [arch...@mail-archive.com]
-=-=-=-=-=-=-=-=-=-=-=-

Re: [vpp-dev] Generic flow type for 5G flow enhancement

[Xu, Ting]

Hi, all

This patch for generic flow for 5G enhancement has been submitted for a long 
time, but not reviewed yet. Could you please help have a look at it and give 
some comments?

https://gerrit.fd.io/r/c/vpp/+/34802

Another patch that adding generic flow type to VAPI was also submitted just now.

https://gerrit.fd.io/r/c/vpp/+/35614/1

Thanks for your time to review!

Best Regards,
Xu Ting

From: Xu, Ting
Sent: Wednesday, March 2, 2022 9:55 AM
To: vpp-dev@lists.fd.io
Cc: dmar...@me.com
Subject: Generic flow type for 5G flow enhancement

Hi, all

We propose a new flow type, generic flow, in patch 
https://gerrit.fd.io/r/c/vpp/+/34802.

The generic flow is based on DDP (Dynamic Device Personalization) function and 
Parser Emulator module (which is supported in DPDK now and planned to be 
enabled in VPP native AVF). Using generic flow pattern, we do not need to 
create new packet and field type and offset in API parser (VAPI or VPP CLI) for 
every new protocols. We can create flows for any protocol immediately as long 
as supported by DDP. The generic flow can be used to support 5G related 
protocols in different scenarios, such as GTPU PDU-type and QFI.
The input of generic flow are two binary strings. They may be not intuitive to 
users, so we will also upstream one tool named PacketForge later to help create 
a user friendly interface, using json configuration file or Scapy format 
command.

The patch was sent several weeks ago and it passed all CI check. Could anyone 
help review? Thanks.

Best Regards,
Xu Ting

-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#21016): https://lists.fd.io/g/vpp-dev/message/21016
Mute This Topic: https://lists.fd.io/mt/89492952/21656
Group Owner: vpp-dev+ow...@lists.fd.io
Unsubscribe: https://lists.fd.io/g/vpp-dev/unsub [arch...@mail-archive.com]
-=-=-=-=-=-=-=-=-=-=-=-

Re: [vpp-dev] Segmentation fault when dpdk number-rx-queues > 1 in startup.conf

[Xu, Ting]

f6e6e242 in vlib_main (vm=0x7fff76800680, input=0x7fff6f66ffa8) 
at /root/networking.dataplane.fdio.vpp/src/vlib/main.c:2017
#28 0x76ed02ce in thread0 (arg=140735181489792) at 
/root/networking.dataplane.fdio.vpp/src/vlib/unix/main.c:671
#29 0x76d0ebf8 in clib_calljmp () at 
/root/networking.dataplane.fdio.vpp/src/vppinfra/longjmp.S:123
#30 0x7fffc9f0 in ?? ()
#31 0x76ecfdfe in vlib_unix_main (argc=59, argv=0x446500) at 
/root/networking.dataplane.fdio.vpp/src/vlib/unix/main.c:751
#32 0x00406b23 in main (argc=59, argv=0x446500) at 
/root/networking.dataplane.fdio.vpp/src/vpp/vnet/main.c:342

Please tell me if any more info needed

Best Regards,
Xu Ting

> -Original Message-
> From: Damjan Marion 
> Sent: Friday, March 4, 2022 9:13 PM
> To: Xu, Ting 
> Cc: vpp-dev@lists.fd.io
> Subject: Re: [vpp-dev] Segmentation fault when dpdk number-rx-queues > 1
> in startup.conf
> 
> 
> Dear Xu Ting,
> 
> Data you provided is not sufficient to help you.
> i.e. providing backtrace may help us understand where problem is.
> 
> —
> Damjan
> 
> 
> 
> > On 03.03.2022., at 08:10, Xu, Ting  wrote:
> >
> > Hi, all
> >
> > I meet one issue that when I set dpdk rx queue number larger than 1
> (which also enables RSS) in startup.conf, for example:
> >
> > dev default {
> >   # Number of receive queues, enables RSS
> >   # Default is 1
> >   num-rx-queues 2
> > }
> >
> > When start VPP, it will meet segmentation fault, the error log is:
> >
> > ……
> > dpdk [debug ]: [0] interface dpdk_eth0 created
> > interface/rx-queue [debug ]: set_input_node: node dpdk-input for
> interface dpdk_eth0
> > interface/rx-queue [debug ]: register: interface dpdk_eth0 queue-id 0
> thread 1
> > interface/rx-queue [debug ]: register: interface dpdk_eth0 queue-id 1
> thread 2
> > dpdk [debug ]: [0] configuring device name: :d8:00.0, numa: 1, driver:
> net_ice, bus: pci
> > dpdk [debug ]: [0] Supported RX offloads: vlan-strip ipv4-cksum udp-cksum
> tcp-cksum qinq-strip
> > outer-ipv4-cksum vlan-filter vlan-extend scatter
> > timestamp keep-crc rss-hash
> > dpdk [debug ]: [0] Configured RX offloads: ipv4-cksum scatter
> > dpdk [debug ]: [0] Supported TX offloads: vlan-insert ipv4-cksum udp-
> cksum tcp-cksum sctp-cksum
> > tcp-tso outer-ipv4-cksum qinq-insert multi-segs
> > mbuf-fast-free outer-udp-cksum
> > dpdk [debug ]: [0] Configured TX offloads: ipv4-cksum udp-cksum tcp-
> cksum multi-segs
> > Segmentation fault (core dumped)
> >
> > I think I find the bad commit:
> ce4083ce48958d9d3956e8317445a5552780af1a (“dpdk: offloads cleanup”)
> > Does anyone also meet issue? Is there any solution to it? Thanks!
> >
> > Best Regards,
> > Xu Ting
> >
> > 

#0  0x7fff71cdacea in ice_sq_send_cmd_nolock (hw=0x0, cq=0x0, desc=0x0, 
buf=0x0, buf_size=0, cd=0x0)
at ../src-dpdk/drivers/net/ice/base/ice_controlq.c:939
#1  0x7fff71cfe88b in ice_sq_send_cmd (hw=0x7fd2bf7b7e80, 
cq=0x7fd2bf7b9940, desc=0x7fff6ca760f0, buf=0x7fe2c0252d40, buf_size=6, cd=0x0)
at ../src-dpdk/drivers/net/ice/base/ice_controlq.c:1126
#2  0x7fff71d2e744 in ice_sq_send_cmd_retry (hw=0x7fd2bf7b7e80, 
cq=0x7fd2bf7b9940, desc=0x7fff6ca760f0, buf=0x7fe2c0252d40, buf_size=6, cd=0x0)
at ../src-dpdk/drivers/net/ice/base/ice_common.c:1729
#3  0x7fff71d3c90f in ice_aq_send_cmd (hw=0x7fd2bf7b7e80, 
desc=0x7fff6ca760f0, buf=0x7fe2c0252d40, buf_size=6, cd=0x0)
at ../src-dpdk/drivers/net/ice/base/ice_common.c:1788
#4  0x7fff71d3d3b8 in ice_aq_alloc_free_res (hw=0x7fd2bf7b7e80, 
num_entries=1, buf=0x7fe2c0252d40, buf_size=6, opc=ice_aqc_opc_alloc_res, 
cd=0x0)
at ../src-dpdk/drivers/net/ice/base/ice_common.c:2124
#5  0x7fff71d3d4dd in ice_alloc_hw_res (hw=0x7fd2bf7b7e80, type=96, num=1, 
btm=false, res=0x7fff6ca78602)
at ../src-dpdk/drivers/net/ice/base/ice_common.c:2154
#6  0x7fff71ee4fa7 in ice_alloc_prof_id (hw=0x7fd2bf7b7e80, 
blk=ICE_BLK_RSS, prof_id=0x7fff6ca7866a "\247l")
at ../src-dpdk/drivers/net/ice/base/ice_flex_pipe.c:3310
#7  0x7fff71f05a04 in ice_add_prof (hw=0x7fd2bf7b7e80, blk=ICE_BLK_RSS, 
id=17179875328, ptypes=0x7fe2c0253c3c "", attr=0x0, attr_cnt=0,
es=0x7fe2c0253b10, masks=0x7fe2c0253bda, fd_swap=true) at 
../src-dpdk/drivers/net/ice/base/ice_flex_pipe.c:5029
#8  0x7fff71f222f5 in ice_flow_add_prof_sync (hw=0x7fd2bf7b7e80, 
blk=ICE_BLK_RSS, dir=ICE_FLOW_RX, prof_id=17179875328, segs=0x7fe2c0253d40,
segs_cnt=1 '\001', acts=0x0, acts_cnt=0 '\000', prof=0x7fff6ca7d178) at 
../src-dpdk/drivers/net/ice/base/ice_flow.c:2245
#9  0x7fff71f2319e in ice_flow_add_prof (hw=0x7fd2bf7b7e80, 
blk=ICE_BLK_RSS, dir=ICE_FLOW_RX,

[vpp-dev] Segmentation fault when dpdk number-rx-queues > 1 in startup.conf

[Xu, Ting]

Hi, all

I meet one issue that when I set dpdk rx queue number larger than 1 (which also 
enables RSS) in startup.conf, for example:

dev default {
  # Number of receive queues, enables RSS
  # Default is 1
  num-rx-queues 2
}

When start VPP, it will meet segmentation fault, the error log is:

..
dpdk [debug ]: [0] interface dpdk_eth0 created
interface/rx-queue [debug ]: set_input_node: node dpdk-input for interface 
dpdk_eth0
interface/rx-queue [debug ]: register: interface dpdk_eth0 queue-id 0 thread 1
interface/rx-queue [debug ]: register: interface dpdk_eth0 queue-id 1 thread 2
dpdk [debug ]: [0] configuring device name: :d8:00.0, numa: 1, driver: 
net_ice, bus: pci
dpdk [debug ]: [0] Supported RX offloads: vlan-strip ipv4-cksum udp-cksum 
tcp-cksum qinq-strip
outer-ipv4-cksum vlan-filter vlan-extend scatter
timestamp keep-crc rss-hash
dpdk [debug ]: [0] Configured RX offloads: ipv4-cksum scatter
dpdk [debug ]: [0] Supported TX offloads: vlan-insert ipv4-cksum udp-cksum 
tcp-cksum sctp-cksum
tcp-tso outer-ipv4-cksum qinq-insert multi-segs
mbuf-fast-free outer-udp-cksum
dpdk [debug ]: [0] Configured TX offloads: ipv4-cksum udp-cksum tcp-cksum 
multi-segs
Segmentation fault (core dumped)

I think I find the bad commit: ce4083ce48958d9d3956e8317445a5552780af1a ("dpdk: 
offloads cleanup")
Does anyone also meet issue? Is there any solution to it? Thanks!

Best Regards,
Xu Ting

-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#20946): https://lists.fd.io/g/vpp-dev/message/20946
Mute This Topic: https://lists.fd.io/mt/89520993/21656
Group Owner: vpp-dev+ow...@lists.fd.io
Unsubscribe: https://lists.fd.io/g/vpp-dev/unsub [arch...@mail-archive.com]
-=-=-=-=-=-=-=-=-=-=-=-

[vpp-dev] Generic flow type for 5G flow enhancement

[Xu, Ting]

Hi, all

We propose a new flow type, generic flow, in patch 
https://gerrit.fd.io/r/c/vpp/+/34802.

The generic flow is based on DDP (Dynamic Device Personalization) function and 
Parser Emulator module (which is supported in DPDK now and planned to be 
enabled in VPP native AVF). Using generic flow pattern, we do not need to 
create new packet and field type and offset in API parser (VAPI or VPP CLI) for 
every new protocols. We can create flows for any protocol immediately as long 
as supported by DDP. The generic flow can be used to support 5G related 
protocols in different scenarios, such as GTPU PDU-type and QFI.
The input of generic flow are two binary strings. They may be not intuitive to 
users, so we will also upstream one tool named PacketForge later to help create 
a user friendly interface, using json configuration file or Scapy format 
command.

The patch was sent several weeks ago and it passed all CI check. Could anyone 
help review? Thanks.

Best Regards,
Xu Ting

-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#20937): https://lists.fd.io/g/vpp-dev/message/20937
Mute This Topic: https://lists.fd.io/mt/89492952/21656
Group Owner: vpp-dev+ow...@lists.fd.io
Unsubscribe: https://lists.fd.io/g/vpp-dev/unsub [arch...@mail-archive.com]
-=-=-=-=-=-=-=-=-=-=-=-