Troubles using pdump to capture the packets

Troubles using pdump to capture the packets

[廖書華]

Dear all,

In our environment, the DPDK version is for 19.11.
Currently, we try to use pdump to capture the packet of our own
application, however, we got the below error in our application and pdump
respectively. (If we replace our application with test-pmd, pdump will not
get the error)

Error in our application :
EAL: failed to parse device “vdev:net_pcap_rx_0”
EAL: Failed to hotplug add device on primary
Error in pdump : The error is caused since function *rte_dev_probe()* will
return -14
EAL: Failed to hotplug add device
EAL: Error - exiting with code: 1
Cause: vdev creation failed:create_mp_ring_vdev:700

Taking the reference from DPDK email list, I already added the library
(-Wl,-lrte_pdump -Wl,-lrte_pmd_pcap -Wl,-lpcap) in my application Makefile
and recompile my application again, however, still got the same error
message. Can someone give me some suggestions ?

*Here's the complete log related to DPDK on my application.*
[oran@localhost app]$ sudo -E LD_LIBRARY_PATH=$LD_LIBRARY_PATH
./run_o_ru.sh.........
.
.
.
xran_ethdi_init_dpdk_io: Calling rte_eal_init:wls_1 -c 0x800008001 -n2
--iova-mode=pa --socket-mem=8192 --socket-limit=8192 --proc-type=auto
--file-prefix wls_1 -w 0000:00:00.0 EAL: Detected 96 lcore(s) EAL: Detected
2 NUMA nodes EAL: Auto-detected process type: PRIMARY EAL: Multi-process
socket /var/run/dpdk/wls_1/mp_socket EAL: Selected IOVA mode 'PA' EAL:
Probing VFIO support... EAL: VFIO support initialized dump : 0 EAL: PCI
device 0000:18:0a.0 on NUMA socket 0 EAL: probe driver: 8086:154c
net_i40e_vf EAL: using IOMMU type 1 (Type 1) initializing port 0 for TX,
drv=net_i40e_vf Port 0 MAC: 00 11 22 33 44 55 Port 0: nb_rxd 4096 nb_txd
4096 Checking link status portid [0] ... done Port 0 Link Up - speed 10000
Mbps - full-duplex EAL: PCI device 0000:18:0a.1 on NUMA socket 0 EAL: probe
driver: 8086:154c net_i40e_vf initializing port 1 for TX, drv=net_i40e_vf
Port 1 MAC: 00 11 22 33 44 55 Port 1: nb_rxd 4096 nb_txd 4096 EAL: failed
to parse device "vdev:net_pcap_rx_0" EAL: Failed to hotplug add device on
primary Checking link status portid [1] ... done Port 1 Link Up - speed
10000 Mbps - full-duplex vf 0 local SRC MAC: 00 11 22 33 44 55 vf 0 remote
DST MAC: 00 11 22 33 44 66 vf 1 local SRC MAC: 00 11 22 33 44 55 vf 1
remote DST MAC: 00 11 22 33 44 66

*Here's the complete log of pdump.*
[oran@localhost pdump]$ sudo ./dpdk-pdump -- --pdump
'port=0,queue=*,tx-dev=/tmp/tx.pcap,rx-dev=/tmp/rx.pcap' EAL: Detected 96
lcore(s) EAL: Detected 2 NUMA nodes EAL: Multi-process socket
/var/run/dpdk/wls_1/mp_socket_10639_24929237def08 EAL: Selected IOVA mode
'VA' EAL: Probing VFIO support... EAL: VFIO support initialized EAL: PCI
device 0000:00:04.0 on NUMA socket 0 EAL: probe driver: 8086:2021
rawdev_ioat EAL: PCI device 0000:00:04.1 on NUMA socket 0 EAL: probe
driver: 8086:2021 rawdev_ioat EAL: PCI device 0000:00:04.2 on NUMA socket 0
EAL: probe driver: 8086:2021 rawdev_ioat EAL: PCI device 0000:00:04.3 on
NUMA socket 0 EAL: probe driver: 8086:2021 rawdev_ioat EAL: PCI device
0000:00:04.4 on NUMA socket 0 EAL: probe driver: 8086:2021 rawdev_ioat EAL:
PCI device 0000:00:04.5 on NUMA socket 0 EAL: probe driver: 8086:2021
rawdev_ioat EAL: PCI device 0000:00:04.6 on NUMA socket 0 EAL: probe
driver: 8086:2021 rawdev_ioat EAL: PCI device 0000:00:04.7 on NUMA socket 0
EAL: probe driver: 8086:2021 rawdev_ioat EAL: PCI device 0000:18:00.0 on
NUMA socket 0 EAL: probe driver: 8086:1572 net_i40e EAL: PCI device
0000:18:00.1 on NUMA socket 0 EAL: probe driver: 8086:1572 net_i40e EAL:
PCI device 0000:18:0a.0 on NUMA socket 0 EAL: probe driver: 8086:154c
net_i40e_vf EAL: using IOMMU type 1 (Type 1) EAL: PCI device 0000:18:0a.1
on NUMA socket 0 EAL: probe driver: 8086:154c net_i40e_vf EAL: PCI device
0000:3d:00.0 on NUMA socket 0 EAL: probe driver: 8086:37d2 net_i40e EAL:
PCI device 0000:3d:00.1 on NUMA socket 0 EAL: probe driver: 8086:37d2
net_i40e EAL: PCI device 0000:80:04.0 on NUMA socket 1 EAL: probe driver:
8086:2021 rawdev_ioat EAL: PCI device 0000:80:04.1 on NUMA socket 1 EAL:
probe driver: 8086:2021 rawdev_ioat EAL: PCI device 0000:80:04.2 on NUMA
socket 1 EAL: probe driver: 8086:2021 rawdev_ioat EAL: PCI device
0000:80:04.3 on NUMA socket 1 EAL: probe driver: 8086:2021 rawdev_ioat EAL:
PCI device 0000:80:04.4 on NUMA socket 1 EAL: probe driver: 8086:2021
rawdev_ioat EAL: PCI device 0000:80:04.5 on NUMA socket 1 EAL: probe
driver: 8086:2021 rawdev_ioat EAL: PCI device 0000:80:04.6 on NUMA socket 1
EAL: probe driver: 8086:2021 rawdev_ioat EAL: PCI device 0000:80:04.7 on
NUMA socket 1 EAL: probe driver: 8086:2021 rawdev_ioat EAL: Failed to
hotplug add device EAL: Error - exiting with code: 1 Cause: vdev creation
failed:create_mp_ring_vdev:700

Best Regards,
Shu-hua, Liao

Re: Troubles using pdump to capture the packets

[Stephen Hemminger]

On Wed, 13 Oct 2021 14:10:17 +0800
廖書華 <sim860927@gmail.com> wrote:

> Dear all,
> 
> In our environment, the DPDK version is for 19.11.
> Currently, we try to use pdump to capture the packet of our own
> application, however, we got the below error in our application and pdump
> respectively. (If we replace our application with test-pmd, pdump will not
> get the error)
> 
> Error in our application :
> EAL: failed to parse device “vdev:net_pcap_rx_0”

The pdump support requires the pcap PMD to be enabled in the build.
What is your DPDK config?

Re: Troubles using pdump to capture the packets

[廖書華]

Dear Stephen,

Thanks for your response !!!
Yes, I already set *CONFIG_RTE_LIBRTE_PMD_PCAP=y *and
*CONFIG_RTE_LIBRTE_PDUMP=y *in the file "dpdk-19.11/config/common_base"
then build DPDK.
Also, in the files "dpdk-19.11/x86_64-native-linuxapp-icc/.config" and
"dpdk-19.11/x86_64-native-linuxapp-icc/.config.orig", they also show that
*CONFIG_RTE_LIBRTE_PMD_PCAP=y *and *CONFIG_RTE_LIBRTE_PDUMP=y.*

It seems that it already enabled pcap PMD of DPDK.
Do you have other suggestions ?


Best Regards,
Shu-hua, Liao


Stephen Hemminger <stephen@networkplumber.org> 於 2021年10月14日 週四 上午12:55寫道：

> On Wed, 13 Oct 2021 14:10:17 +0800
> 廖書華 <sim860927@gmail.com> wrote:
>
> > Dear all,
> >
> > In our environment, the DPDK version is for 19.11.
> > Currently, we try to use pdump to capture the packet of our own
> > application, however, we got the below error in our application and pdump
> > respectively. (If we replace our application with test-pmd, pdump will
> not
> > get the error)
> >
> > Error in our application :
> > EAL: failed to parse device “vdev:net_pcap_rx_0”
>
> The pdump support requires the pcap PMD to be enabled in the build.
> What is your DPDK config?
>

RE: Troubles using pdump to capture the packets

[Pattan, Reshma]

> -----Original Message-----
> From: 廖書華 <sim860927@gmail.com>


> Yes, I already set *CONFIG_RTE_LIBRTE_PMD_PCAP=y *and
> *CONFIG_RTE_LIBRTE_PDUMP=y *in the file "dpdk-
> 19.11/config/common_base"
> then build DPDK.
> Also, in the files "dpdk-19.11/x86_64-native-linuxapp-icc/.config" and "dpdk-
> 19.11/x86_64-native-linuxapp-icc/.config.orig", they also show that
> *CONFIG_RTE_LIBRTE_PMD_PCAP=y *and *CONFIG_RTE_LIBRTE_PDUMP=y.*
> 
> It seems that it already enabled pcap PMD of DPDK.
> Do you have other suggestions ?

Hi,

Few options you can double check
1)Make sure your primary application is calling rte_pdump_init()/ret_pudmp_uninit() to initialize/uninitialize the pdump library.
2)If you are using a shared library build,   double check you are  linking pcap pmd properly in primary build as explained in below link
https://www.mail-archive.com/users@dpdk.org/msg05039.html
https://stackoverflow.com/questions/62795017/dpdk-pdump-failed-to-hotplug-add-device
3)If you are passing any pci device using  eal "-w" option to primary,  try to pass the same device to secondary also using "-w" option . 

If you still see the issue please paste the full primary and secondary application run log with command that you are running.
Also what kind of build you are using.

Thanks,
Reshma
  

Re: Troubles using pdump to capture the packets

[廖書華]

[-- Attachment #1.1: Type: text/plain, Size: 4520 bytes --]

Dear Reshma,

Thanks for your kind response ! Unfortunately, it still can't work.

1) I double checked the return value (int) of these two functions, both of
them return 0.
2) About the DPDK building, we are not using the shared library build.
[image: Screenshot from 2021-10-15 10-19-29.png]
The DPDK library is for static (.a), by the way, for the primary
application we used, I already added static library (librte_pdump.a,
librte_pmd_pcap.a) in the Makefile (
https://drive.google.com/file/d/1Y2Oua1T41U86qPUiiOtw7_XCIJzqJDmP/view?usp=sharing
).
[image: Screenshot from 2021-10-15 10-29-18.png]
3) For the primary side, I saw the log showed "-w 0000:00:00.0", which are
the input of function rte_eal_init(). Therefore, I also tried to parse "-w
0000:00:00.0" on pdump side, however, it still got the error.
*- Log in pdump side*
https://drive.google.com/file/d/1QIHRr4A-4lh3QvfQUF8CLwJkutY23OPO/view?usp=sharing

*- Log in primary side (Red font I think is related to DPDK)*
https://drive.google.com/file/d/1_ItKwrDebaA1skuei8RExkftudUlXiAS/view?usp=sharing

---------------------

P.S 1. On primary side, If we change  "-w 0000:00:00.0" to  "-w
0000:86:02.0 -w 0000:86:02.1" which is the two bus-info we bind for DPDK.
[image: Screenshot from 2021-10-15 11-52-53.png]
We will get the error  *EAL: Failed to attach device on primary process, * do
you have any idea what is the root cause of this issue ?
xran_ethdi_init_dpdk_io: Calling rte_eal_init:wls_0 -c 0x2000021 -n2
--iova-mode=pa --socket-mem=8192 --socket-limit=8192 --proc-type=auto
--file-prefix wls_0 -w 0000:86:02.0 -w 0000:86:02.1 EAL: Detected 40
lcore(s) EAL: Detected 2 NUMA nodes EAL: Auto-detected process type:
PRIMARY EAL: Multi-process socket /var/run/dpdk/wls_0/mp_socket EAL:
Selected IOVA mode 'PA' EAL: Probing VFIO support... EAL: VFIO support
initialized EAL: PCI device 0000:86:02.0 on NUMA socket 1 EAL: probe
driver: 8086:154c net_i40e_vf EAL: using IOMMU type 1 (Type 1) EAL: PCI
device 0000:86:02.1 on NUMA socket 1 EAL: probe driver: 8086:154c
net_i40e_vf Return value (int) of rte_pdump_init() : 0 EAL: PCI device
0000:86:02.0 on NUMA socket 1 EAL: Failed to attach device on primary
process initializing port 0 for TX, drv=net_i40e_vf Port 0 MAC: 00 11 22 33
44 66 Port 0: nb_rxd 4096 nb_txd 4096 Checking link status portid [0] ...
done Port 0 Link Up - speed 10000 Mbps - full-duplex EAL: PCI device
0000:86:02.1 on NUMA socket 1 EAL: Failed to attach device on primary
process initializing port 1 for TX, drv=net_i40e_vf Port 1 MAC: 00 11 22 33
44 66 Port 1: nb_rxd 4096 nb_txd 4096 Checking link status portid [1] ...
done Port 1 Link Up - speed 10000 Mbps - full-duplex vf 0 local SRC MAC: 00
11 22 33 44 66 vf 0 remote DST MAC: 00 11 22 33 44 55 vf 1 local SRC MAC:
00 11 22 33 44 66 vf 1 remote DST MAC: 00 11 22 33 44 55

P.S. 2 : For the DPDK environment we to build, we select
x86_64-native-linuxapp-icc and Insert VFIO module.

Best Regards,
Shu-hua, Liao


Pattan, Reshma <reshma.pattan@intel.com> 於 2021年10月14日 週四 下午5:52寫道：

>
>
> > -----Original Message-----
> > From: 廖書華 <sim860927@gmail.com>
>
>
> > Yes, I already set *CONFIG_RTE_LIBRTE_PMD_PCAP=y *and
> > *CONFIG_RTE_LIBRTE_PDUMP=y *in the file "dpdk-
> > 19.11/config/common_base"
> > then build DPDK.
> > Also, in the files "dpdk-19.11/x86_64-native-linuxapp-icc/.config" and
> "dpdk-
> > 19.11/x86_64-native-linuxapp-icc/.config.orig", they also show that
> > *CONFIG_RTE_LIBRTE_PMD_PCAP=y *and *CONFIG_RTE_LIBRTE_PDUMP=y.*
> >
> > It seems that it already enabled pcap PMD of DPDK.
> > Do you have other suggestions ?
>
> Hi,
>
> Few options you can double check
> 1)Make sure your primary application is calling
> rte_pdump_init()/ret_pudmp_uninit() to initialize/uninitialize the pdump
> library.
> 2)If you are using a shared library build,   double check you are  linking
> pcap pmd properly in primary build as explained in below link
> https://www.mail-archive.com/users@dpdk.org/msg05039.html
>
> https://stackoverflow.com/questions/62795017/dpdk-pdump-failed-to-hotplug-add-device
> 3)If you are passing any pci device using  eal "-w" option to primary,
> try to pass the same device to secondary also using "-w" option .
>
> If you still see the issue please paste the full primary and secondary
> application run log with command that you are running.
> Also what kind of build you are using.
>
> Thanks,
> Reshma
>
>

[-- Attachment #1.2: Type: text/html, Size: 11413 bytes --]

[-- Attachment #2: Screenshot from 2021-10-15 10-19-29.png --]
[-- Type: image/png, Size: 5900 bytes --]

[-- Attachment #3: Screenshot from 2021-10-15 10-29-18.png --]
[-- Type: image/png, Size: 262649 bytes --]

[-- Attachment #4: Screenshot from 2021-10-15 11-52-53.png --]
[-- Type: image/png, Size: 157006 bytes --]

RE: Troubles using pdump to capture the packets

[Pattan, Reshma]

From: 廖書華 <sim860927@gmail.com>
Sent: Friday, October 15, 2021 5:16 AM
To: Pattan, Reshma <reshma.pattan@intel.com>
Cc: Stephen Hemminger <stephen@networkplumber.org>; users@dpdk.org; 林庭安 <lingwanjae@gmail.com>
Subject: Re: Troubles using pdump to capture the packets

>>Dear Reshma,

>>Thanks for your kind response ! Unfortunately, it still can't work.


Also, could you please paste me the lbpcap-devel package version on your system.

Thanks,
Reshma

Pattan, Reshma <reshma.pattan@intel.com<mailto:reshma.pattan@intel.com>> 於 2021年10月14日 週四 下午5:52寫道：


> -----Original Message-----
> From: 廖書華 <sim860927@gmail.com<mailto:sim860927@gmail.com>>


> Yes, I already set *CONFIG_RTE_LIBRTE_PMD_PCAP=y *and
> *CONFIG_RTE_LIBRTE_PDUMP=y *in the file "dpdk-
> 19.11/config/common_base"
> then build DPDK.
> Also, in the files "dpdk-19.11/x86_64-native-linuxapp-icc/.config" and "dpdk-
> 19.11/x86_64-native-linuxapp-icc/.config.orig", they also show that
> *CONFIG_RTE_LIBRTE_PMD_PCAP=y *and *CONFIG_RTE_LIBRTE_PDUMP=y.*
>
> It seems that it already enabled pcap PMD of DPDK.
> Do you have other suggestions ?

Hi,

Few options you can double check
1)Make sure your primary application is calling rte_pdump_init()/ret_pudmp_uninit() to initialize/uninitialize the pdump library.
2)If you are using a shared library build,   double check you are  linking pcap pmd properly in primary build as explained in below link
https://www.mail-archive.com/users@dpdk.org/msg05039.html
https://stackoverflow.com/questions/62795017/dpdk-pdump-failed-to-hotplug-add-device
3)If you are passing any pci device using  eal "-w" option to primary,  try to pass the same device to secondary also using "-w" option .

If you still see the issue please paste the full primary and secondary application run log with command that you are running.
Also what kind of build you are using.

Thanks,
Reshma