Hi,
I’m trying to compile gst-nvarguscamerasrc from source because I need to modify the hard-coded gain and exposure time ranges. I cloned the code from https://nv-tegra.nvidia.com/r/tegra/gst-src/gst-nvarguscamera.git and checkouted tag jetson_35.3.1.
To make sure that this build process works, I first ran make and built a new libgstnvarguscamerasrc.so without changing any code. Then I proceeded to put this new .so file at /usr/lib/aarch64-linux-gnu/gstreamer-1.0/libgstnvarguscamerasrc.so. My Jetson Linux version is R35.3.1.
I ran the following command to capture and save a test image:
gst-launch-1.0 nvarguscamerasrc sensor-id=0 num-buffers=1 ! "video/x-raw(memory:NVMM), format=(string)NV12, width=(int)3840, height=(int)2160" ! nvvidconv ! pngenc ! filesink location=test1.png
Then I encountered this error:
This command worked before I replaced the libgstnvarguscamerasrc.so
I noticed that the md5sum of the original libgstnvarguscamerasrc.so is identical to the one shipped in Linux_for_Tegra/nv_tegra/nv_sample_apps/nvgstapps.tbz2 after nvgstapps.tbz2 is untarred by apply_binaries.sh. Note: this Linux_for_Tegra comes from the driver package (Jetson_Linux_R35.3.1_aarch64.tbz2) downloaded from the office website.
The new libgstnvarguscamerasrc.so that I built however has a different checksum and it’s much bigger in size.
I have also compared the gst-nvarguscamerasrc source code (r35.3.1) that I cloned with the source code in gst-nvarguscamera_src.tbz2 from the Driver Package source (public_sources.tbz2 for 35.3.1) downloaded from the website. They are identical.
Another thing I have tried was building libgstnvarguscamerasrc.so for R35.1 from the source cloned from https://nv-tegra.nvidia.com/r/tegra/gst-src/gst-nvarguscamera.git and placing it onto a Jetson with r35.1. The gst capture command worked fine.
So is there a bug in the R35.3.1 gst-nvarguscamera source code?
Thanks