Error when installing MM API using JetPack / Not able to make the device visible from host

Hi,
I want to use MM API to access hardware encoder on Jetson TX2. So I use JetPack to install this API on my Jetson. But I obtain the following error :
Can someone help me to know what it is the issue ?
Regards,

Hi Nostalgie,
It looks to be mismatch of revision of system and tegra_multimedia_api sample. Do you use Jetpack to re-flash the system also?

This is duplicate of https://devtalk.nvidia.com/default/topic/1031835/jetson-tx2/using-mm-api-with-jetson-tx2/

Hi Danell,

Thanks for your reply. I use JetPack only to install MM API and other components without flashing the system. For the other thread, I have created it before installing JetPack and obtaining errors.

Regards,

Hi Danell,

As you said there is a mismatch of revision of system and samples. Now I should flash the target also. But when following the steps in the documentation and before flashing it, I can not make the Jetson TX2 visible by my host PC. Note that I use Virtual Machine ubuntu 16.04 (on windows 10) as host machine. Is the VM causes this problem ?

Regards.

Most VM issues are related to USB problems (and sometimes ethernet problems). Version mismatch is more likely from using a newer or older JetPack versus the version used with the Jetson’s base install. You can see the version of L4T on the Jetson via:

head -n 1 /etc/nv_tegra_release

If R28.1, then JetPack3.1…if R28.2, then JetPack3.2.

FYI, I suspect a VM is less of a problem for extra package install than it is for flash because USB is not used for the extra package install. Only ethernet is used for package installs. Even so there may be a few VM “gotchas”.

Hi linuxdev,

Thanks for your reply. My problem is resolved by flashing my TX2 through a native linux. It was important to flash it because I have the R27.0.1 revision.

Regards,

But I have another question, if you can help me please.

  1. Where we can find details about using the hardware encoder and decoder on Jetson TX2 using MM API. Is the documentation is limited to the samples in tegra-multimedia-api folder ? Because in the MM API Reference site : file:///home/nvidia/NVIDIA_Tegra_Linux_MultimediaAPIReference/nvl4t_docs/l4t_mm_test_group.html, there isn’t an explication/ description for 00-video-decode and 01-video-encode codes. Can u give me a simple description for them ?

  2. In the following thread : https://devtalk.nvidia.com/default/topic/1024356/solved-harware-accelerated-video-encoding-from-c-buffer-with-gstreamer-omxh264enc-and-filesink/, It is said that MM API can be used to encode on Jetson TX2 using H264 hardware encoder but using Video4Linux2 as the data source, so we can not send our proper data (from external application or from buffers) to the encoder when using MM API ?

Regards.

That’s a topic I have not been through. What I’d recommend doing to start with is to go here:
https://developer.nvidia.com/embedded/downloads
…type “multimedia” in the search box.

Some people have extended examples on public git repositories, though I couldn’t tell you about any specific one.