• Content Count

  • Joined

  • Last visited

Everything posted by FGuerzoni

  1. Thank you, 'your' was for shortness I hoped it could support at least main profile. I have to encode 1080p streams and baseline profile it's not well suited. Regards F
  2. Hi @Ubobrov, thank you very much for all the infos. Very very interesting. About H.264 encoding, what profile types your driver supports? (base, main, high, ...)
  3. After a fresh new complete rebuild and reinstall following a clean procedure I got the same troubles. Giving up for now until I get some grip on the issue. Currently it's just a black box for me. @jernej Thank you for all your support and patience.
  4. I was not aware of that option. My bad. I only had to left out --enable-libopencv because even if installed it cannot be seen. Don't know why. Anyway, after ffmpeg and mpv rebuilt anything changed. > ffmpeg ffmpeg version n4.0.4-0ubuntu0.18.04.1 Copyright (c) 2000-2019 the FFmpeg developers built with gcc 7 (Ubuntu/Linaro 7.4.0-1ubuntu1~18.04.1) configuration: --prefix=/usr --extra-version=0ubuntu0.18.04.1 --toolchain=hardened --libdir=/usr/lib/arm-linux-gnueabihf --incdir=/usr/include/arm-linux-gnueabihf --enable-gpl --disable-stripping --enable-avresample --enable-avisyn
  5. Thank you, I did a first quick attempt because of little time available, but I was not able to build ffmpeg with so many options. I got troubles installing so many libraries and make them available to ffmpeg. I'll retry later.
  6. After complete rebuild and reinstall of everything it's still not working. It falls back to software decoding with cpu over 350% The file to decode is the standard Big Bucket Bunny at 1080p > whereis libdrm libdrm: /usr/lib/arm-linux-gnueabihf/libdrm.so /usr/include/libdrm /usr/share/libdrm > whereis libudev libudev: /lib/arm-linux-gnueabihf/libudev.so /usr/include/libudev.h /usr/share/man/man3/libudev.3.gz > ls /usr/src/linux-headers-5.3.1-sunxi/ Documentation Makefile arch certs drivers include ipc lib net scripts sound usr Kconfig Mod
  7. I checked the armbian config-default.conf where the option is empty instead of "yes" INSTALL_HEADERS="" # install kernel headers package Tomorrow I'll rebuild everything and I'll let you know if it's working Thanks and regards
  8. Maybe I have problems on kernel headers. Please confirm me that this is the right way to search for them > whereis libdrm libdrm: /usr/lib/arm-linux-gnueabihf/libdrm.so /usr/include/libdrm /usr/share/libdrm > whereis libudev libudev: /lib/arm-linux-gnueabihf/libudev.so /usr/include/libudev.h /usr/share/man/man3/libudev.3.gz > sudo ls /usr/src/kernels/$(uname -r) ls: cannot access '/usr/src/kernels/5.3.0-sunxi': No such file or directory
  9. You are right, even if I don't understand why because I've explicitly set '--enable-v4l2-request' > ffmpeg ffmpeg version n4.0.4 Copyright (c) 2000-2019 the FFmpeg developers built with gcc 7 (Ubuntu/Linaro 7.4.0-1ubuntu1~18.04.1) configuration: --enable-v4l2-request --enable-libudev --enable-libdrm libavutil 56. 14.100 / 56. 14.100 libavcodec 58. 18.100 / 58. 18.100 libavformat 58. 12.100 / 58. 12.100 libavdevice 58. 3.100 / 58. 3.100 libavfilter 7. 16.100 / 7. 16.100 libswscale 5. 1.100 / 5. 1.100 libswresample 3. 1.100 /
  10. I had some progresses using 'null'. But it's logging that it's using software decoding. From that the fact the cpu is very high. > mpv --vo=null --hwdec=auto -v bbb_1080p.mp4 [cplayer] Command line options: '--vo=null' '--hwdec=auto' '-v' 'bbb_1080p.mp4' [cplayer] mpv 0.28.2 (C) 2000-2017 mpv/MPlayer/mplayer2 projects [cplayer] built on Sun Sep 22 11:14:34 UTC 2019 [cplayer] ffmpeg library versions: [cplayer] libavutil 56.14.100 [cplayer] libavcodec 58.18.100 [cplayer] libavformat 58.12.100 [cplayer] libswscale 5.1.100 [cplayer] libavfilter
  11. Yes I have /dev/dri/card0 > ls /dev/dri/card0 /dev/dri/card0 but I don't have any monitor connected. NanoPiNeo doesn't have any HDMI output. Sorry, I wasn't aware that the physical monitor is mandatory. So with my device the only way is to dig into ffmpeg and understand why it's apparently not using the hw accels. I'll post results in case I'll get them. Thank you again Regards F
  12. Thank you jernej, there's no desktop environment at all and I'm running command from the command line. I'm using just the standard megous DEV branch patched with the kernel patches you suggested. I'm already able to run on the same device the VAAPI stack, which is quite limited because it's able to correctly decode only the base profile. From previous posts I got the fact that the ffmpeg way could decode any h264 profiles. Is it right? If so I'm very interested to make it work, because I don't have any idea when the VAAPI stack will be fully implemented.
  13. Digging a bit more Calling v4l2-ctl --log-status I'm getting the same 'Inappropriate ioctl for device' seen before > v4l2-ctl --list-devices --verbose cedrus (platform:cedrus): /dev/video0 VIDIOC_QUERYCAP: ok > v4l2-ctl --log-status --verbose VIDIOC_QUERYCAP: ok VIDIOC_LOG_STATUS: failed: Inappropriate ioctl for device
  14. Sure, thank you. Read below. What is not clear to me is if your patches still use VAAPI. In that case I've not yet installed (because of compilation errors to solve) libva-v4l2-request as suggested by https://linux-sunxi.org/Sunxi-cedrus libva-v4l2-request The main userspace component that supports the Cedrus VPU driver is the libva-v4l2-request VAAPI backend. It is available at: Repository: https://github.com/bootlin/libva-v4l2-request In the very same page it's also suggested for VLC, but could be the same for FFMPEG, to export LIBVA_DRIVER_NAME=v4l2_r
  15. Thank you for quick reply. That was first attempt I made but I got troubles: > mpv --hwdec=auto --vo=drm ipc47_seq01.mp4 Playing: ipc47_seq01.mp4 (+) Video --vid=1 (*) (h264 1920x1088 9.939fps) [vo/drm] VT_GETMODE failed: Inappropriate ioctl for device [vo/drm] Failed to set up VT switcher. Terminal switching will be unavailable. [vo/drm] Cannot retrieve DRM resources: Operation not supported [vo/drm] Failed to create KMS. Error opening/initializing the selected video_out (--vo) device. Video: no video No video or audio streams selected. Exiting..
  16. Thank you jernej for the detailed instructions. I successfully built 5.3 kernel with cedrus driver enabled and the patched ffmpeg 4.0.4 on a NanoPiNeo I'm interested in processing streams from ip cameras on headless devices and I don't require mpv. I tried to directly run ffmpeg to test hardware decoding but it doesn't seem to take advantage of the hardware acceleration. CPU is still very high. I piped the output to a named pipe in order to discard it to remove any writing overhead to SD disk. Have you got any hints to take advantage of the hw accel? > cat outpipe &g
  17. I really have poor experience with libva-v4l2-request library. Maybe you could find useful that Sunxi-Cedrus linux release-2019.03 is a patched version of 5.0.7 kernel release About libva-v4l2-request library, I tried to compile it. And yes I too had to disable version check. And I found that the version from github master was easier to compile (less issues) that the one suggested by the sunxi wiki, which was the release-2019.03
  18. Thank you very much. I missed CONFIG_VIDEO_SUNXI_CEDRUS. After doing the image rebuild I tried the v4l2-request-test. And both MPEG2 and H.264 baseline decoding work fine. Just for reference, with nanopi neo (H3) and 864*480 slices, H.264 I frame is decoded in 6ms and B frames are decoded in less than 2ms. The only issue I found is about video_buffers data arrays utilization. With 16 frames (the default) it seems that only the first frame of 16 is correctly stored in arrays. It seems that arrays with index>0 contain either dirty or no data at all. Nee
  19. @drice how did you manage to get /dev/video0 working as required by v4l2-request-test? Without it's impossible to have the buffers source. regards F
  20. Thank you very much for the tip. I built the image without any particular issue and did the boot: Welcome to Ubuntu Bionic with Armbian Linux 5.3.0-rc3-sunxi System load: 0.14 0.24 0.11 Up time: 3 min Memory usage: 14 % of 491MB IP: CPU temp: 42°C Usage of /: 6% of 15G I can start to check if h.264 baseline decoder is working with the current image Regards F
  21. I tried to do same thing some months ago without success working with nanopi neo. I had problems applying the patches (manually derived from https://github.com/bootlin/linux-cedrus [release-2019.03] ) to armbian dev because I'm quite new to armbian and kernel stuff. But recently I noticed that both mainline next and 5.3-rc3 added h.264 decoding (baseline profile): I really cannot manage to point armbian build system to either 5.3-rc3 or next-20190808. I think also that should be important to check (from https://linux-sunxi.org/Sunxi-cedrus): So if anyone could m
  22. I managed to boot ARMBIAN 5.83 and kernel 5.0.9 patched with cedrus stuff. It has been quite tricky because cedrus march release referred to kernel 5.0.1 and I didn't find any source related to latest kernel to diff and patch. Then I flashed a SD card with full image Armbian_5.83_Nanopineo_Ubuntu_bionic_dev_5.0.9.img I don't know if it's ok because I still have to move to userspace and build the test app. Anyway: what sounds strange to me is that on the target device I still have linux header dev files that seem not patched. For exampe videodev2.h does not conta
  23. Thank you very much for the hint. I'll try it and I'll update the post with results. Regards F.
  24. Thank you for quick reply. I didn't want to build with armbian build system because my final goal is to have sunxi-cedrus 5.0.0 image working on nanopi neo. So I downloaded latest 4.19.36 kernel release and build it to test the my build procedure. I'm not aware of standard patches. Regards F
  25. Hi, I have a nanopi neo SBC and my final goal is to get Sunxi-cedrus 2019.03_Release (patched kernel 5.0.0 release) installed on device to test cedrus driver. Did a manual kernel build and tried to dpkg -i *.deb files on the board but the kernel hangs on boot. I read also but that didn't help me so much So I tried same procedure to replace the armbian standard 4.19.20 kernel with an official mainline 4.19.36 version. I found same exact issue so I think I'm missing something. Please have a quick read at the procedure I followed and put any advice to he