Hello,
As the title said, is there any relationship between VPE and DSS?
Thanks
Terence
This thread has been locked.
If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.
Hello,
As the title said, is there any relationship between VPE and DSS?
Thanks
Terence
Hello,
Can you elaborate on what you mean by "relationship"? A hardware relationship or software use case of some sort?
thanks,
Alex
Hi Terence,
There is no relation between VPE and DSS. Both are independent HW modules and can work independently.
What functionality are enabled in VPE and DSS in above diagram?
Rgds,
Brijesh
Hello terence xia, we are expecting your feedback/explanation on what your use case is. Generally, you should be able to feed display output from camera. There is no direct HW link, but it is doable via software (you may try the demos in the SDK for example)
Hi Alex,
You can find our use case on the below ticket, my colleague have described it.
On the other hand, if we remove VPE link, and enable DSP1 & DSP2 core, the same problem will occurs. We set "PROC_DSP1_INCLUDE=yes PROC_DSP2_INCLUDE=yes" and "status =”okay” " in dra7-evm-infoadas.dts to enable dsp in the cfg.mk.
Thanks
Terence
Hi Alex,
Sorry. For above message "on the other hand", I'm not describe it very clear. I will describe it again here,
"On the other hand, if we remove VPE link, and enable DSP1 & DSP2 core, also it will occurs problem. And only enable DSP1, it is normal. And if remove DSS link, and enable DSP1 & DSP2 ,it works normal. We set "PROC_DSP1_INCLUDE=yes PROC_DSP2_INCLUDE=yes" and "status =”okay” " in dra7-evm-infoadas.dts to enable dsp in the cfg.mk."
Add log about enable DSP2 , and disbale DSP2 in the attachment.
Thanks
Hi Alex,
To enable DSP1 & DSP2 , it may be another problem. So, I create a new ticket.

Thanks for the details. We will ask experts for help here with your initial question, they will reply here soon.
Good you submitted a new thread for the different issue, that's the right approach.
Hi Alex,
Thank you very much.
In this ticket,for more clear, I describe it again here, our problem is VPE link and DSS link can not work together, it will occur PVK error. If we remove DSS link, it will become normal.On the other hand, if we remove VPE link, also it is normal. Below is our usecase, and please find the attachment.
Hi
Error edma 43300000.edma: CCERR 0x00000001
indicates that System dma's Q0 is full and it has cross the watermakr/threshold value.
I am not sure if this is due to the impact of the l3_noc error or this itself is an issue.
Can you please get these readings when this issue happens?
omapconf read 0x43300600
omapconf read 0x43300604
omapconf read 0x43300620
Thanks
RamPrasad
Hi Ramprasad,
According to your suggestion, we try to reproduce this issue many times. But we can not find the PVR_K error in the log any more, and l3_noc error still remain. The system's running time may be not enough. I will try again,and response to you soon. Below is the omapconf reading. Please find the newest log in the attachment.
Thanks
terence
============================================
root@dra7xx-evm:~# omapconf read 0x43300600
00000000
root@dra7xx-evm:~# omapconf read 0x43300604
00000000
root@dra7xx-evm:~# omapconf read 0x43300620
00001010
Hi Terence,
The EDMA readings were required if the issue is reproduced.
Please share the logs when issue is reproduced again and edma readings.
Thanks
RamPrasad
Hi Terence,
omapconf read 0x43300600
omapconf read 0x43300604
omapconf read 0x43300620
These commands are edma reading. I wanted these when issue was reproduced.
Please share these when issue gets reproduced.
Thanks
RamPrasad
Hi RamPrasad,
We had reproduced the issue. While, it can not execute omapconf command ant more. The system work abnormal, it always print "PVR_K" error.
Is there any way to hardcode the command of edma reading?
Please find the newest log in the attachment.
Thanks
Terence
Hi Terence,
The new log also shows that SGX-HW recovery happenned due to accessing an invalid address.
omap_l3_noc error confirms this. But this log is not giving any information which address is wrong, it requires some more tracing.
This issue is reproduced over very long duration execution of the usecase(over a week).
Is there a way for you reproduce issue as early as possible?
Thanks
RamPrasad
Hi RamPrasad,
1. Which way can get the more log you want?
2.For "Is there a way for you reproduce issue as early as possible?"
Yes, we can reproduce it easy now. We try 10 times, and it will reproduce 3 times. Because, we enable QT to send more data into the usecase.
Thanks
Terence