V4l2-compliance err VIDIOC_G/S_PARAM:FAILD on TX2.Use OH01A10 camera

Hi everyone.
I got an error when develop driver on TX2.My camera is OH01A10.
When I use v4l2-compliance , I got one ‘Failed’: VIDIOC_G/S PARAM failed.
When I use v4l2 to get raw data.I got log:
vi-output, ov56-9248 [000] … 374.793831: tegra_channel_capture_setup: vnc_id 0 W 1280 H 800 fmt 20
vi-output, ov56-9248 [000] … 374.793869: tegra_channel_capture_frame: sof:-549620698956.-269759139712
Meanwhile,the dmesg tells :
[ 284.152478] tegra-vi4 15700000.vi: PXL_SOF syncpt timeout! err = -11
[ 284.158938] tegra-vi4 15700000.vi: tegra_channel_error_recovery: attempting to reset the capture channel
THe attachment are the whole log.
What show I do new,thanks for your help.

dmesg -v4l2.txt (72.4 KB)
trace-v4l2.txt (564.4 KB)

hello 798971767,

this is expected v4l2 compliance test failure since there’s no implementation. you may ignore this test item.

however, the reported timestamp looks incorrect, please double check your sensor streaming.
please also review the sensor pixel clock settings, thanks

Thanks for you reply.
May I ask, is there any other parm in dts or in driver file could cause this error?
Other blogs show that mclk_khz, active_w, active_h are influential as well, is there any other than these.
If we make some mistake in gain parma of exp time, will they cause this error?

hello 798971767,

it’s an invalidate capture frame. it’s VI engine to wait camera start-of-frame, and program the timestamp to capture buffer.
is it possible to probe the MIPI signaling? please check you’ve correct signal sending to CSI brick.
BTW, may I also know which JetPack release version you’re using?

Hi JerryChang,
Thanks for your reply. We use Sdkmanager to root the TX2, the JetPack version in it is 4.6.2.
May I ask , if the quality of the camera patch cord is not good, will it cause this error?

hello Benson.Gao,

this kind of failure usually due to sensor signal, it’s from hardware side.
the software configuration are device tree settings, which correspond to sensor hardware settings.

tegra186-camera-e3326-a00.dtsi (9.3 KB)
Hello JerryChang,
This is our dtsi file. Can you find any error in it or can you tell me which param cound cause this error?
Thanks a lot.

hello Benson.Gao,

is your camera OH01A10 really use CSI-C interface?
please review Sensor Pixel Clock for your pixel_clk_hz settings.

it looks you mostly based-on e3326 reference driver to have implementation.
there’s string devname = "ov5693 2-0036"; to detect driver sources, did you also modify ov5693.c driver to adapt with your camera hardware?
please also revise this string, badge = "e3326_front_P5V27C"; otherwise it’ll load ov5693 camera calibration files.

Hi JerryChang,
Thank you for your reply.
I will review the interface.I have seen Sensor Pixel Click setting.I see there are three ways to calc the pixel_clock.
And I get three result.So I use the biggest one, is that right?

As for the second question. It is right I based-on e3326 reference driver to have implementation. And I have change the param and function in ov5693 driver files including one ov5693.c and four ov5693.h and one table.

yes, it’s okay.

you should at least revise this string, badge = "e3326_front_P5V27C";
it’s pre-built binary to parse this unique name for identifies camera module. please see-also badge property in sensor programming guide.

Hi JerryChang,
Thanks for your fast reply.
I will change this value and try again.
Wish you a good day

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.