Video stream capture doesn't work because of discarding frame error, err_data 131072

[mi-hana]

I’m trying to get video stream from Sony IMX390 camera connected to Jetson AGX Orin 64GB Developer Kit (35.4.1).

What I wrote in my application code is to repeat buffer dequeue/queue100 times with…
DequeueBuffer → ioctl(fd_, VIDIOC_DQBUF, buffer)
QueueBuffer → ioctl(fd_, VIDIOC_QBUF, buffer)

I successfully captured the first eight frames (maybe because I had set buffer size at eight) , however, ioctl(fd_, VIDIOC_DQBUF, buffer) didn’t work/respond at ninth time then my code stacked.

dmesg implies something had been going wrong even during the first eight frame captures as below.

[ 159.522918] tegra-camrtc-capture-vi tegra-capture-vi: corr_err: discarding frame 0, flags: 0, err_data 131072
[ 159.556064] tegra-camrtc-capture-vi tegra-capture-vi: corr_err: discarding frame 0, flags: 0, err_data 131072
[ 159.589407] tegra-camrtc-capture-vi tegra-capture-vi: corr_err: discarding frame 0, flags: 0, err_data 131072
[ 159.622744] tegra-camrtc-capture-vi tegra-capture-vi: corr_err: discarding frame 0, flags: 0, err_data 131072
[ 159.656055] tegra-camrtc-capture-vi tegra-capture-vi: corr_err: discarding frame 0, flags: 0, err_data 131072
[ 159.689399] tegra-camrtc-capture-vi tegra-capture-vi: corr_err: discarding frame 0, flags: 0, err_data 131072
[ 159.722709] tegra-camrtc-capture-vi tegra-capture-vi: corr_err: discarding frame 0, flags: 0, err_data 131072
[ 159.756043] tegra-camrtc-capture-vi tegra-capture-vi: corr_err: discarding frame 0, flags: 0, err_data 131072

Could you tell me what “err_data 131072” mean?
And also I’d appreciate it if you could tell me a next step to try.

[JerryChang]

hello mi-hana,

@mi-hana

I successfully captured the first eight frames

may I know what’s your test pipelines?

@mi-hana

discarding frame 0, flags: 0, err_data 131072

those messages discarding frame it’s an error reported by VI.
it means an unsuccessfully capture state, it has to restart the capture engine to restore the failure.

BTW,
there’re some camera bug fixes, you may refer to Release Notes (r35.5.0) ,
is it possible for moving to the latest JP-5.1.3/l4t-r35.5.0 for verification?

[mi-hana]

Hi, Jerry. Thanks for your quick reply.

may I know what’s your test pipelines?
Start → Queue buffer 8 times → Dequeue → Queue → Dequeue → Queue…

After the post, I found specified image format in my code was wrong; my sensor provided images in 12-bit depth however I had set kPixelFormatNV16 in my code.
By correcting it to kPixelFormatSRGGB12, all the errors were cleared and the code successfully worked.

I can close this thread here. Thanks also for information regarding the new JP/l4t release.