CSI camera driver Timeout with SPURIOUS_DATA_STREAM_4 on the TX2

Autonomous Machines Jetson & Embedded Systems Jetson TX2
#1

We have the CSI driver capturing frames about 10 frames per second but it’s configured for 60 frames per second.
We can see that it continues to time out every couple of frames, we can see the timeout on dmesg.

We were able to get additional information from capturing the trace as they did on this post: https://devtalk.nvidia.com/default/topic/1029929/csi-camera-driver-syncpt-timeout-error/?offset=8

We were wondering what does the SPURIOUS_DATA_STREAM_4 flag from CSIMUX_STREAM mean ?
In our trace we get the following continuously logging :

kworker/0:0-4     [000] ...1   125.453414: rtcpu_vinotify_handle_msg: tstamp:4252947092 tag:CSIMUX_STREAM channel:0xff frame:0 vi_tstamp:4252946665 data:0x00010000
kworker/0:0-4     [000] ...1   125.453418: rtcpu_vinotify_handle_msg: tstamp:4253468457 tag:CSIMUX_STREAM channel:0xff frame:0 vi_tstamp:4253468029 data:0x00010000
kworker/0:0-4     [000] ...1   125.453420: rtcpu_vinotify_handle_msg: tstamp:4253989857 tag:CSIMUX_STREAM channel:0xff frame:0 vi_tstamp:4253989428 data:0x00010000
kworker/0:0-4     [000] ...1   125.453422: rtcpu_vinotify_handle_msg: tstamp:4254511208 tag:CSIMUX_STREAM channel:0xff frame:0 vi_tstamp:4254510781 data:0x00010000

We tried my things but nothing seems to help, any ideas of what the issue can be?

#2

@nicolas.norena
Spurious data means VI sees some other packets before FS packet, as VI alwasy expect the 1st packet to be frame start

#3

Hi @ShaneCCC, I am having the same issue. Do you have any idea what could be the reason for this?
@nicolas.norena did you solve the problem? What caused it?

#4

@a7az0th
It’s could be the sensor output some maybe user defined data before the Frame Start.
You may need to check with the sensor vendor to check if can disable it to try.

#5

I was never able to resolved this issue, We ended up going on a different direction. If I recall correctly, the driver supported the basic configuration of the chip’s registers. I ran out of ideas, and we had to go take a different direction.

It would be worth looking at what ShanneCCC said.

If you are able to figure it out, I am curious to know how you solve it.

good luck.

#6