[00:09] *** BrianG61UK_ has quit IRC (Quit: Leaving)
[00:09] *** BrianG61UK has quit IRC (Quit: Leaving)
[00:25] *** BrianG61UK has joined #linux-media
[00:27] *** BrianG61UK_ has joined #linux-media
[01:08] *** svarbanov has quit IRC (Ping timeout: 480 seconds)
[01:17] *** svarbanov has joined #linux-media
[01:24] *** b-rad has quit IRC (Read error: Connection timed out)
[01:58] *** camus has joined #linux-media
[02:31] *** camus has quit IRC (Remote host closed the connection)
[02:31] *** camus has joined #linux-media
[03:31] *** johfjeld has quit IRC (Ping timeout: 480 seconds)
[03:39] *** johfjeld has joined #linux-media
[04:18] *** NiksDev has joined #linux-media
[04:31] *** camus1 has joined #linux-media
[04:31] *** camus has quit IRC (Read error: Connection reset by peer)
[05:36] *** jagan_ has joined #linux-media
[06:04] *** camus1 has quit IRC (Remote host closed the connection)
[06:04] *** camus has joined #linux-media
[07:07] *** ao2 has joined #linux-media
[07:17] *** jmondi has quit IRC (Ping timeout: 480 seconds)
[07:35] <hverkuil> tfiga: ping
[07:44] *** camus1 has joined #linux-media
[07:44] *** jmondi has joined #linux-media
[07:48] *** camus has quit IRC (Ping timeout: 480 seconds)
[08:10] <tfiga> hverkuil: pong
[08:21] <hverkuil> tfiga: can you review Sergey's non-coherent v4 series? I'm almost done with my review, I'm hoping v5 will be the final version.
[08:21] <hverkuil> But I'd like your comment (if any) for v4 as well.
[08:26] <tfiga> hverkuil: sure, I'll do it
[08:27] <tfiga> I was waiting for you to take a look first, since you had some comments last time
[08:27] <hverkuil> I should be done soon.
[08:39] *** johfjeld has quit IRC (Remote host closed the connection)
[08:41] *** jagan_ has quit IRC (Remote host closed the connection)
[10:16] <hverkuil> tfiga: posted my last review of Sergey's series.
[10:49] *** camus has joined #linux-media
[10:54] *** camus1 has quit IRC (Ping timeout: 480 seconds)
[10:56] <tfiga> hverkuil: got it, I'll take a look
[11:00] *** dafna2[m] has quit IRC (Quit: Bridge terminating on SIGTERM)
[11:00] *** tomba has quit IRC (Quit: Bridge terminating on SIGTERM)
[11:00] *** gnurou has quit IRC (Quit: Bridge terminating on SIGTERM)
[11:17] *** bud_4_hill[m] has joined #linux-media
[12:10] <mort_> well, I'm stumped. I'm now actually perfectly implementing the m2m API from what I can tell, but now I'm not even getting a V4L2_EVENT_SOURCE_CHANGE event
[12:20] *** camus1 has joined #linux-media
[12:24] *** camus has quit IRC (Ping timeout: 480 seconds)
[12:26] <mort_> oh, nevermind. This makes sense.
[12:39] *** jm_h has joined #linux-media
[13:28] *** camus has joined #linux-media
[13:29] *** camus1 has quit IRC (Remote host closed the connection)
[13:38] <ezequielg> hverkuil: ping
[13:38] *** jagan_ has joined #linux-media
[13:39] *** jernej_ has joined #linux-media
[13:39] *** jernej has quit IRC (Quit: Free ZNC ~ Powered by LunarBNC: https://LunarBNC.net)
[13:40] *** BrianG61UK_ has quit IRC (Remote host closed the connection)
[13:42] *** BrianG61UK_ has joined #linux-media
[13:50] *** camus1 has joined #linux-media
[13:51] *** camus has quit IRC (Read error: Connection reset by peer)
[14:40] *** dafna2[m] has joined #linux-media
[14:41] *** gnurou has joined #linux-media
[14:41] *** tomba[m] has joined #linux-media
[14:53] *** jernej_ is now known as jernej
[14:55] <mort_> Are there any docs anywhere on what sort of buffers a particular driver requires
[14:55] <mort_> I know I've asked before, but I don't think I got a clear yes or now
[14:56] <mort_> the venus decoder for example doesn't work if I feed it NAL units separately, it needs the first buffer to contain both the SPS and the PPS at least
[14:56] <ezequielg> mort_: https://01.org/linuxgraphics/gfx-docs/drm/media/uapi/v4l/dev-decoder.html
[14:57] <ezequielg> well actually kernel.org/doc/html/latest/userspace-api/media/v4l/dev-decoder.html
[14:57] <mort_> ezequielg: which part of that documents what exactly the buffers you send have to contain?
[14:57] <ezequielg> this should be standardized, and so you can use ffmpeg or gstreamer on any driver.
[14:58] <ezequielg> i.e. i.MX6 CODA works out of the box. 
[14:58] <ezequielg> "A stateful video decoder takes complete chunks of the bytestream (e.g. Annex-B H.264/HEVC stream, raw VP8/9 stream)"
[14:58] <mort_> where is the standard
[14:58] <mort_> what kind of chunking
[14:58] <mort_> you can't just send individual NAL units, that doesn't work
[14:59] <hverkuil> ezequielg: the problem mort_ is encountering is that some older drivers (venus) predating that spec do not fully comply. I.e. you can't just pass the full compressed stream in the buffers, they need to be pre-parsed. That's not documented, svarbanov might know the details. (At least, this is how I understand it).
[14:59] <hverkuil> ezequielg: pong
[14:59] <ezequielg> ah, if venus is not as per-spec, I don't know.
[15:00] <mort_> I think it's relatively per spec these days
[15:00] <mort_> more than it used to at least
[15:00] <ezequielg> at least CODA works out of the box in gstreamer, and AFAIK gstreamer implements that spec, and doesn't (or shouldn't) do anything hw specific.
[15:00] <mort_> the issue I'm encountering now is that I can decode my test h264 bitstream perfectly fine..... for 103 frames, after which the poll times out
[15:01] <ezequielg> hverkuil: hi there! summer over? :)
[15:02] <hverkuil> haven't been on vacation yet, it's actually been quite busy (mentoring interns).
[15:03] <ezequielg> hverkuil: ouch.... :) so we are about to submit the VP9 support, in addition to RKVDEC, the patchset will include Hantro G2. We have tested on i.MX8MQ, i.MX8MP and RK3399. The drivers are passing a significant amount of test vectors in fluster conformance test suite, and the VP9 control interface has been reviewed mulitple times by people here at Collabora. Having said that, I don't think it's 5.15 material.
[15:03] <ezequielg> (cont) I'd like you and Nicolas to review it carefully.
[15:03] <ezequielg> and Nicolas is on hoidays for 2 weeks.
[15:04] <ezequielg> So, just keep that in mind. It will be submitted soon (this week I hope), but we are aiming at 5.16.
[15:04] <mort_> alright, new strategy: I'm gonna try to build a wrapper around gstreamer's API which matches video decoder interface of my existing decoder
[15:04] <hverkuil> OK, I agree, 5.15 is too optimistic :-)
[15:05] <ezequielg> I'd like to have the Tiled YUV pixel format cleanup for 5.15, if possible.
[15:05] <ezequielg> If you like the cleanup, that is :)
[15:05] <ezequielg> Not an issue if that goes to 5.16 as well, in fact.
[15:05] <ezequielg> On our side, we are prioritizing VP9 over HEVC, just FYI.
[15:06] <hverkuil> that makes sense (vp9 over hevc)
[15:06] <hverkuil> I'll see if I can review the tiled formats tomorrow.
[15:12] <KitsuWhooa> hverkuil: Apologies, but since you're around, any chances on this getting merged before 5.14 is released? https://patchwork.kernel.org/project/linux-media/list/?series=505605
[15:12] <KitsuWhooa> And would replying with another Tested-by help at all?
[15:13] <hverkuil> syoung: didn't you deal with that regression? ^^^^
[15:21] <jernej> hverkuil, ezequielg: can be at least HEVC scaling matrix control patches merged?
[15:21] <jernej> it's working for sure, at least on Cedrus
[15:22] <hverkuil> ezequielg: I'm waiting for a Ack from you for those scaling patches.
[15:38] *** jm_h has quit IRC (Ping timeout: 480 seconds)
[15:42] <ezequielg> hverkuil: oh, yes. I was planning to jernej.
[15:42] <ezequielg> note that we need to better spec the expectation for that control, but i can go on a separate patch, to avoid delaying things.
[15:42] <ezequielg> jernej: do you have hevc cedrus patches queued?
[15:43] <ezequielg> or pending to be sent? (just curious)
[15:43] <jernej> they are on ML for some time
[15:44] <ezequielg> cedrus?
[15:46] <jernej> yes
[15:49] *** jm_h has joined #linux-media
[16:10] *** BigBadBeautifulBoris has quit IRC (Remote host closed the connection)
[16:15] *** BigBadBeautifulBoris has joined #linux-media
[16:36] *** BrianG61UK has quit IRC (Read error: Connection reset by peer)
[16:58] *** ao2 has quit IRC (Remote host closed the connection)
[17:31] *** camus has joined #linux-media
[17:37] *** camus1 has quit IRC (Ping timeout: 480 seconds)
[17:53] *** jagan_ has quit IRC (Remote host closed the connection)
[18:26] *** BrianG61UK has joined #linux-media
[19:44] *** montjoie_ is now known as montjoie
[19:46] *** lexano has quit IRC (Ping timeout: 480 seconds)
[19:55] *** lexano has joined #linux-media
[20:07] *** camus1 has joined #linux-media
[20:10] *** camus has quit IRC (Remote host closed the connection)
[21:32] *** djrscally has joined #linux-media
[21:38] *** BigBadBeautifulBoris has quit IRC (Remote host closed the connection)
[21:48] *** jm_h has quit IRC (Quit: Leaving)
[22:26] *** NiksDev has quit IRC (Ping timeout: 480 seconds)
[23:04] *** djrscally has quit IRC (Quit: Konversation terminated!)
[23:33] *** svarbanov has quit IRC (Ping timeout: 480 seconds)