Re : nvh264sldec + CUDA

"christophe.lafolet--- via gstreamer-devel" <[email protected]>
Newsgroups gmane.comp.video.gstreamer.devel
Message-ID <-m4fh-sv9ij4-jeg4l8h3dfbw-lgk7hw-rau9ayj6sxaultqycg19naju-bxrmsx-j64nsoczy3e2-lssnfh-18ydwn-297qxe6epope-hzcv7rim6dmqtbmsb-cajzfk-4cptwutwk1vl-oby4ov-xy34i8.1686054338611@email.android.com>
It seems my problem comes from the caps negotiation (gst_nv_decoder_negotiate) which is called several times.
The first time, after requesting gst.cuda.context :

as I can't implement allocation query (my application is in Java) , a default pool is created with CUDAMemory and some buffers start using this pool.

The next time, after requesting gst.gl.app_context : a new pool is created with GLMemory.

Is it a voluntary behaviour or a bug ?

Christophe

-------- Message original --------
De : Christophe Lafolet via gstreamer-devel <[email protected]>
Date : ven. 2 juin 2023 à 20:24
À : [email protected]
Cc : Christophe Lafolet <[email protected]>
Objet : nvh264sldec + CUDA

hello

I try to decode a H264 video with nvh264sldec

my appsink accept only video/x-raw and video/x-raw(memory:GLMemory) caps
: no CUDAMemory in my caps

on the need_context msg, I set only my GL wrapped context : no CUDA
context set

I receive several buffers with CUDAMemory (that I can't use) and after
others buffers with GLMemory

Christophe

Environment : CentOS7, Gstreamer 1.22.3, NVIDIA RTX A4000
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.