camus has quit [Read error: Connection reset by peer]
camus has joined #tegra
camus1 has joined #tegra
camus has quit [Remote host closed the connection]
camus1 is now known as camus
camus has quit [Ping timeout: 256 seconds]
camus has joined #tegra
<kwizart>
hello #tegra and happy new year to all the team!
camus1 has joined #tegra
<kwizart>
digetx, I've tested current grate from this morning and it looks very interesting, I wonder if v4l2-m2m is fully wired with tegra124, but at least I've dropped this message "kernel: Failed to attached device 54200000.dc to IOMMU_mapping" I'm not sure why, maybe it was moving to CONFIG_IOMMU_DEFAULT_DMA_STRICT=y instead of LAZY. With this message dropped, I'm not experiencing any tearing under gnome-shell anymore (since IOMMU is working with tegra_drm)
<kwizart>
also nouveau seems to work fine with this grate update !
camus has quit [Read error: Connection reset by peer]
camus1 is now known as camus
<kwizart>
hum, there is no /dev/tegra_vde or /dev/video so I guess not all pieces are there for tegra124...
<kwizart>
tested vdpau on tegra20(paz00) works, but I need to force the -vo=vdpau in mpv or the video doesn't display
<kwizart>
not managed to test v4l2-m2m there yet
camus1 has joined #tegra
camus has quit [Read error: Connection reset by peer]
<digetx>
I haven't pushed out vdpau-tegra and gstreamer v4l patches yet, still testing and working on them
<digetx>
and I already worked around all the troubles, at least vdpau-tegra will work okay using V4L
<digetx>
can't say that I'm positive about V4L, it adds 2k lines of code to VDE driver to do the same thing that driver did without V4L, on top of it I'm not impressed about that stateless V4L API is made on top of the very old-fashioned core V4L API which has design problems and creates troubles for userspace; V4L memory management also upsets, I don't understand how anyone could seriously use it; on the positive side, it all should be solvable