Jetson TX2 PXL_SOF syncpt timeout! error with custom camera

Hello experts.

I have issue with our custom camera.
This module has isp and working with i2c command and mipi.
Basically, It working without issue but when we command downscale(for example 1/2) using i2c, it is not working with below error.
I don’t know how to debug it. because mipi & vi generate error while image data receiving and finish it, I can’t get any data for debug and analyze.

Could you check my error message and share tips for debug?(Is there any way to get dump data in vi?)

[ 449.051614] tegra-vi4 Status: 2 channel:00 frame:0FA1
[ 449.057890] tegra-vi4 timestamp sof 461630651008 eof 461655828544 data 0x000000a0
[ 449.067285] tegra-vi4 capture_id 13 stream 0 vchan 0
[ 450.035224] tegra-vi4 PXL_SOF syncpt timeout! err = -11
[ 450.035236] tegra-vi4 ATOMP_FE syncpt timeout! err = -11
[ 451.059136] tegra-vi4 ATOMP_FE syncpt timeout!

[Debug Trace]
kworker/0:0-2188 [000] …1 401.930020: rtos_queue_peek_from_isr_failed: tstamp:12952889219 queue:0x0b4a3c58
kworker/0:2-250 [000] …1 449.043044: rtos_queue_peek_from_isr_failed: tstamp:14425198861 queue:0x0b4a3c58
kworker/0:2-250 [000] …1 449.043048: rtcpu_start: tstamp:14425199803
kworker/0:2-250 [000] …1 449.043050: rtcpu_vinotify_handle_msg: tstamp:14425291044 tag:CSIMUX_STREAM channel:0xff frame:0 vi_tstamp:1540388716 data:0x00000001
kworker/0:2-250 [000] …1 449.043051: rtcpu_vinotify_handle_msg: tstamp:14425958386 tag:CHANSEL_PXL_SOF channel:0x00 frame:4001 vi_tstamp:1541055956 data:0x00000001
kworker/0:2-250 [000] …1 449.043052: rtcpu_vinotify_handle_msg: tstamp:14425958536 tag:ATOMP_FS channel:0x00 frame:4001 vi_tstamp:1541055970 data:0x00000000
kworker/0:2-250 [000] …1 449.043053: rtcpu_vinotify_handle_msg: tstamp:14426270850 tag:CHANSEL_LOAD_FRAMED channel:0x01 frame:4001 vi_tstamp:1541368551 data:0x08000000
kworker/0:2-250 [000] …1 449.095043: rtcpu_vinotify_handle_msg: tstamp:14426745347 tag:CSIMUX_FRAME channel:0x00 frame:4001 vi_tstamp:1541842754 data:0x000000a0
kworker/0:2-250 [000] …1 449.095045: rtcpu_vinotify_handle_msg: tstamp:14426745529 tag:CHANSEL_SHORT_FRAME channel:0x01 frame:4001 vi_tstamp:1541842754 data:0x00000001
kworker/0:2-250 [000] …1 449.095046: rtcpu_vinotify_handle_msg: tstamp:14426745656 tag:ATOMP_FE channel:0x00 frame:4001 vi_tstamp:1541842757 data:0x00000000

The trace show “CHANSEL_SHORT_FRAME” that means csi/vi receive high less than expect.

Could I know received high bytes less than expect?

The sensor output resolution (width x high) not as driver expect.

Yes. I understand.
We set the correct resolution to sensor according to there’s guide.
But csi/vi was failed.(You said because of high data less than expect)
If you are correct, I want prove and notify to sensor company.
So I thought that if we can dump full data in vi/csi then see how many byte exceed then expect, we can prove it.

I can’t dump it sensor side. it only available in nvidia board(csi/vi).
Is there any point to dump or debug it?

Sadly only the trace log for debug.