GoldVip S32G3999A RDB3: Unable to validate NXP example Gateway usecases

[senthilnathan_pas]

1. Followed the instruction of GoldVIP bring up with the S32G3 V 1.12 version.  

2. The RDB3 is booting from the SD card fine however at the end seeing the following error/warning

[ 33.017785] pfeng-slave 46000000.pfe_slave: ERR: (DRIVER) event 1 - Driver runtime error: [pfeng-slave-drv.c:265] PFE controller UP timed out

3. Setup a dcoker container environment as per instructions

4. start the ethernet gateway "A53 slow-path use cases" 

vip@PTCLAPSUBRAS:~/goldvip$ sudo ./eth-slow-path-host.sh -L 3 -d full -t UDP -l 30 enx24f5a2f0fab0 enx002427884bba

Creating nw_ns0 and nw_ns1 network namespaces...

Adding enx24f5a2f0fab0 interface to nw_ns0...

Adding enx002427884bba interface to nw_ns1...

PING 10.0.1.2 (10.0.1.2) 56(84) bytes of data.

--- 10.0.1.2 ping statistics ---

4 packets transmitted, 0 received, 100% packet loss, time 3064ms

Setting IP 192.168.100.1 to enx24f5a2f0fab0 in nw_ns0...

Setting IP 10.0.1.1 to enx002427884bba in nw_ns1...

Checking network between the network namespaces

Failed to connect to the network!

Retrying...

Deleting network namespaces...

Killing iperf3 process...

Creating nw_ns0 and nw_ns1 network namespaces...

Adding enx24f5a2f0fab0 interface to nw_ns0...

Adding enx002427884bba interface to nw_ns1...

connect: Network is unreachable

Setting IP 192.168.100.1 to enx24f5a2f0fab0 in nw_ns0...

Setting IP 10.0.1.1 to enx002427884bba in nw_ns1...

Checking network between the network namespaces

Failed to connect to the network!

Retrying...

Deleting network namespaces...

Killing iperf3 process...

Creating nw_ns0 and nw_ns1 network namespaces...

Adding enx24f5a2f0fab0 interface to nw_ns0...

Adding enx002427884bba interface to nw_ns1...

connect: Network is unreachable

Setting IP 192.168.100.1 to enx24f5a2f0fab0 in nw_ns0...

Setting IP 10.0.1.1 to enx002427884bba in nw_ns1...

Checking network between the network namespaces

Failed to connect to the network!

Retrying...

Deleting network namespaces...

Killing iperf3 process...

Creating nw_ns0 and nw_ns1 network namespaces...

Adding enx24f5a2f0fab0 interface to nw_ns0...

Adding enx002427884bba interface to nw_ns1...

connect: Network is unreachable

Setting IP 192.168.100.1 to enx24f5a2f0fab0 in nw_ns0...

Setting IP 10.0.1.1 to enx002427884bba in nw_ns1...

Checking network between the network namespaces

Failed to connect to the network!

Deleting network namespaces...

Killing iperf3 process...

vip@PTCLAPSUBRAS:~/goldvip$

while the above test is going on from the docker shell, minicom is printing the following….

root@s32g399ardb3:~# /home/root/eth-gw/eth-cleanup-target.sh

Device "pfe0sl" does not exist.

An error occurred in file /home/root/eth-gw/eth-cleanup-target.sh, at line 43

root@s32g399ardb3:~# /home/root/eth-gw/eth-slow-path-target.sh -L 3

Cannot find device "pfe0sl"

An error occurred in file /home/root/eth-gw/eth-slow-path-target.sh, at line 53

root@s32g399ardb3:~# /home/root/eth-gw/eth-cleanup-target.sh

Device "pfe0sl" does not exist.

An error occurred in file /home/root/eth-gw/eth-cleanup-target.sh, at line 43

[alejandro_e]

Hello @senthilnathan_pas , thanks for contacting NXP support.

I will try to replicate the behavior you are seeing to check if there is a problem with the examples provided, it may take a little while to setup all the tools needed and then investigate the root cause, so I appreciate your patience.

I understand that the first error (steps 1 and 2) are from section 5 GoldVIP Quick Start and the second error (steps 3 and 4) are from section 6 Docker of the of the S32G Vehicle Integration Platform (GoldVIP) User Manual for S32G3/UM12007 in my case I have Rev. 1.12.0 — 18 June 2024 (your revision should match, since I'm using GoldVIP 1.12), is my understanding correct?

Regards

[alejandro_e]

Hello @senthilnathan_pas ,

While doing some investigation on the topic I found a very similar case, the explanation to the problem is as follows:

"

GoldVIP includes a multi-instance PFE setup, where the PFE master driver is on the Cortex-M7_0 core (GoldVIP Real-time Gateway Application), and the slave instance runs on the Cortex-A53 clusters (Linux). The master instance has to be up and running, so that the slave one could work.

So, I recommend to follow the steps described in our User Manual and use the NOR Flash boot mode, so that the application running on the Cortex-M7 cores is also loaded and started.

"

The steps to enable the examples are in section 5 GoldVIP Quick Start of the S32G Vehicle Integration Platform (GoldVIP) User Manual for S32G3 [Rev. 1.12.0 — 18 June 2024].

You should see a string similar to "bootloader 0x2 Application 0 image fragment 0 of size 3145728 was loaded from QSPI, address 0x400000."  in the boot logs when the bootloader is configured to boot both cores.

With your current configuration configuration a log like "M7_0 core is not started, ipc-shm-chdev module will not be inserted"  should appear right now, indicating the problem with the Real time core.

Please let me know if this information was useful.