Le mercredi 09 janvier 2019 à 15:28 -0200, Mauro Carvalho Chehab a écrit : > Em Wed, 09 Jan 2019 11:52:45 -0500 > Nicolas Dufresne <nicolas@xxxxxxxxxxxx> escreveu: > > > Le jeudi 10 janvier 2019 à 00:42 +0800, Ayaka a écrit : > > > > There is a UVC media device that supports P010 device. We're discussing > > > > about adding support for it on media. The full thread is at: > > > > > > > > https://lore.kernel.org/linux-media/20190109121900.hbrpttmxz3gaotwx@xxxxxxxxxxxxxxxxxxxxxxxxxx/T/#m8c395156ca0e898e4c8b1e2c6309d912bc414804 > > > > > > > > We've seen that you tried to submit a patch series for DRM adding > > > > support for it at the rockship driver. What's the status of such > > > > series? > > > Rockchip would use another 10bit variant of NV12, which is merged as NV12LE40 at Gstreamer and I sent another patch for it, but I didn’t receive any feedback from that. > > Hmm... We have already NV12 format at media subsystem side: > Documentation/media/uapi/v4l/pixfmt-nv12.rst > > Did we miss some patch from you? NV12 is by definition 8bit per pixel. Rockchip uses a 10bit variant, for which each pixels is packed over 40 bits so that it's bandwidth efficient. It's probably quite complex to handle HW wise. Xilinx also got a variant they are upstream, which is packed over 32 bits, with 2 bits padding. > > > For extra detail, the Rockchip variant is fully packed, with no padding > > bits, while the Microsoft variant uses 16bits per pixels, with 6bits > > padding. It was a mistake to use P010 initially for the Rockchip > > format. > > Yeah, P010 format seems to be a waste of bandwidth. Yet, as we're > now having devices using it (via uvcdriver), it would make sense to > merge the single plane format for it. I would merge the P016 too, as > P010 seems to be just a degradation of it. It's meant for fast CPU indexing, which makes it odd as a streaming format. > > Ayaka, > > Would it be possible for you to re-send the P010/P016 documentation > for single plane, fixing the typo (simliar -> similar)? > > Let's not merge the dual plane formats, as we don't need them > yet at the Kernel. > > Thanks, > Mauro
