narmstrong changed the topic of #linux-amlogic to: Amlogic mainline kernel development discussion - our wiki http://linux-meson.com/ - ml linux-amlogic@lists.infradead.org - Publicly Logged on https://irclog.whitequark.org/linux-amlogic
_whitelogger has joined #linux-amlogic
default__ has joined #linux-amlogic
ldevulder_ has quit [Ping timeout: 252 seconds]
<ndufresne> mjourdan, just tried the jpeg decoder in gst and it didn't work, gstreamer opens the device and enumerate the output format, expecting a list of all possible formats, but your driver will give a subset base on whatever default format was picked
<ndufresne> let's say there is a different interpretation of the spec
<ndufresne> ok, made it work by moving the caps probe after setting the format, it kind of make sense
<ndufresne> we don't strictly need the full list
<ndufresne> mjourdan, though, now gst picks YUV420 by default, this seems broken, maybe it's the KMS side
<ndufresne> ok, YUV420 works from jpeg, but not from mpeg2 decoder
<ndufresne> mjourdan, you'll want to look at the probably, after couple of seeks in gstreamer, a frame get stuck, and I can see this in dmesg
<ndufresne> meson-vdec c8820000.video-decoder: Buffer 4 done but it doesn't exist in m2m_ctx
<lvrp16> narmstrong: do you know if HWCURSOR_API_PLANE is currently supported as well as will be supported?
<lvrp16> for ARMSOC
jakogut_ has joined #linux-amlogic
<ndufresne> mjourdan, ok, added HEVC decoding in gst, HEVC show a colorful X over the image
<ndufresne> quite close though
jakogut_ has quit [Quit: jakogut_]
<ndufresne> mjourdan, ah, that might be a hint, meson-vdec c8820000.video-decoder: Bitstream uses tiles, NotImplemented!
jakogut_ has joined #linux-amlogic
jakogut_ has quit [Read error: Connection reset by peer]
jakogut__ has joined #linux-amlogic
jakogut__ has quit [Quit: jakogut__]
jakogut_ has joined #linux-amlogic
ldevulder_ has joined #linux-amlogic
default__ has quit [Ping timeout: 252 seconds]
vagrantc has joined #linux-amlogic
vagrantc has quit [Quit: leaving]
Amit_T has joined #linux-amlogic
pionen has joined #linux-amlogic
pionen has quit [Client Quit]
chewitt has quit [Quit: Adios!]
ldevulder_ is now known as ldevulder
ccaione has joined #linux-amlogic
<narmstrong> lvrp16: the DRM driver does not export a CURSOR plane, it will need some work to expose it
<lvrp16> narmstrong: is there even a suitable plane for cursor?
<narmstrong> lvrp16: no
<narmstrong> lvrp16: yes
<narmstrong> sorry yes there it, but there is no HW scaler for it, so we must handle the interlaced field switching in irq for each vsync
<narmstrong> this is why it will need some work to add it
<narmstrong> the OSD2 plane is the one
<lvrp16> narmstrong: thanks
<mjourdan> ndufresne: The capture format list that changes based on the selected codec is to comply with https://lkml.org/lkml/2018/7/24/539, see section "Querying capabilties".
<mjourdan> HEVC tiles got fixed like 2 days ago (the branch is force-pushed)
<mjourdan> I'll have a look for YUV420 with other codecs than mjpeg.
<mjourdan> How do you seek with a gstreamer pipeline ?
<mjourdan> ndufresne: Yeah there was a problem with YUV420, I pushed a fix.. Also had to disable it for H.264. Thanks for testing.
<narmstrong> @mjourdan it doesn't work on h.264 ?
<mjourdan> narmstrong: sadly no.. decoder freezes on startup if I don't set the "use 2 planes instead of 3" flag. It's my bad since I hadn't tested it, I saw yuv420p was ok for mjpeg so I just went with it.
<narmstrong> mjourdan: ok
<mjourdan> With the amount of formats and samples, I really need to start automatizing the testing
Amit_T has quit [Quit: Leaving]
Barada has quit [Ping timeout: 240 seconds]
Barada has joined #linux-amlogic
Barada has quit [Ping timeout: 264 seconds]
Barada has joined #linux-amlogic
Barada has quit [Ping timeout: 252 seconds]
vagrantc has joined #linux-amlogic