Multimedia-chat-meeting-2019-01-24

10:06 < pinchartl> welcome to the multimedia meeting !
10:07 < pinchartl> Topic 1. Status Check for the Multimedia Tasks
10:07 < pinchartl> * Jacopo
10:07 < pinchartl> Since last meeting:
10:07 < pinchartl> - adv748x dynamic routing v4 merged
10:07 < pinchartl> - soc_camera removal - small help to finalize removal
10:07 < pinchartl> Until next meeting:
10:07 < pinchartl> - Update vin-tests for the CSVB->TXA use case on E3
10:07 < pinchartl> - Propose a V4L2 API to dynamically negotiate the number of CSI-2 data lanes
10:07 < pinchartl> This should be discussed with Sakari first.
10:07 < pinchartl> - FOSDEM and meetings
10:07 < pinchartl> Issues and blockers: None
10:07 < pinchartl> jmondi: any comment ?
10:07 < jmondi> not really
10:08 < pinchartl> thank you
10:08 < pinchartl> * Kieran
10:08 < pinchartl> Since last meeting:
10:08 < pinchartl> - ADV748x upstreaming/ refactor reset routine, and remote PAGE_WAIT
10:08 < pinchartl> - GMSL v5 review updates, rebase, and fixes for renesas-drivers
10:08 < pinchartl> Until next meeting:
10:08 < pinchartl> - Post GMSL v5
10:08 < pinchartl> - DU D3/E3 support review
10:08 < pinchartl> - Attend PeriPeri Meetings in Brussels and FOSDEM
10:08 < pinchartl> - Continue on Upporting
10:08 < pinchartl> Issues and blockers:
10:08 < pinchartl> - Development laptop failure resulted in a re-install of Ubuntu OS :(
10:08 < pinchartl> kbingham: any comment ?
10:08 < kbingham> none here.
10:09 < pinchartl> thank you
10:09 < pinchartl> * Laurent
10:09 < pinchartl> Since last meeting:
10:09 < pinchartl> - Updated ATF & U-Boot on Salvator-XS M3-N and Draak (was quite painful)
10:09 < pinchartl> - D3/E3 RGB output (complete, pending review)
10:09 < pinchartl> - Reviewed multiplexed streams series
10:09 < pinchartl> Until next meeting:
10:09 < pinchartl> - Skiing holidays
10:09 < pinchartl> - FOSDEM & multimedia meeting
10:09 < pinchartl> Issues and blockers:
10:09 < pinchartl> - Updating ATF and U-Boot was painful
10:09 < pinchartl> Locating the right version of Minimon and the right memory locations is not
10:09 < pinchartl> straightforward. While it is nice to have firmware packages in the wiki for
10:09 < pinchartl> some of the boards, centralizing the information on how to obtain Minimon for
10:09 < pinchartl> each board and where to flash the various components would be useful.
10:09 < pinchartl> any question ?
10:09 < morimoto> Skiing holidays
10:09 < morimoto> where ?
10:10 < pinchartl> very good question :)-
10:10 < pinchartl> https://www.valdisere.com/
10:10 < kbingham> morimoto, We still need a snowboarding holiday! :)
10:10 < pinchartl> in the French Alps
10:11 < morimoto> kbingham: Yeah, I think so
10:11 < morimoto> PeriPeriSki meeting
10:11 < kbingham> :)
10:11 < morimoto> pinchartl: looks nice. Enjoy
10:12 < pinchartl> thank you, I will try to
10:12 < pinchartl> * Morimoto-san
10:12 < pinchartl> Since last meeting:
10:12 < pinchartl> - Investigated ALSA SoC cleanup
10:12 < pinchartl> It is time to final cleanup for ALSA SoC. Initial investigations are done and
10:12 < pinchartl> discussed with the maintainer.
10:12 < pinchartl> Until next meeting:
10:12 < pinchartl> - Continue ALSA SoC work
10:12 < pinchartl> Issues and Blockers: None
10:12 < pinchartl> morimoto: any comment ?
10:12 < morimoto> no comment, thanks
10:12 < pinchartl> thank you
10:12 < pinchartl> * Niklas
10:12 < pinchartl> Since last meeting:
10:12 < pinchartl> - Investigated new LOCKDEP warnings in rcar-vin
10:12 < pinchartl> Had a brief look of new fun LOCKDEP warnings due to the v4l2-async core
10:12 < pinchartl> rework, rcar-vin async callback changes fallout. No clear conclusion so far.
10:12 < pinchartl> Until next meeting:
10:12 < pinchartl> - Try to solve the LOCKDEP warning in rcar-vin
10:12 < pinchartl> - Try to clean up VIN crop and compose code
10:12 < pinchartl> This is needed to to make it easier to merge UDS and PM support.
10:12 < pinchartl> - Attend periperi meetings and FOSDEM.
10:12 < pinchartl> Issues and blockers:
10:12 < pinchartl> - I will be in Brussels from the Jan 29th -- Feb 10th, I expect this
10:12 < pinchartl>   will effect my output.
10:12 < pinchartl> neg: any comment ?
10:13 < neg> No additional comment
10:13 < pinchartl> thank you
10:13 < pinchartl> any other comment or question from anyone regarding the status updates ?
10:15 < pinchartl> Topic 2. Questions from the BSP team
10:15 < pinchartl> - DRM render node support
10:15 < pinchartl> The BSP team was informed by Igalia that R-Car doesn't have DRM render node [1][2]
10:15 < pinchartl> support. Igalia is working on Chromium support, and Chromium uses render nodes to
10:15 < pinchartl> separates the UI process and the GPU rendering for speed-up.
10:15 < pinchartl> [1] https://en.wikipedia.org/wiki/Direct_Rendering_Manager#Render_nodes
10:15 < pinchartl> [2] https://01.org/linuxgraphics/gfx-docs/drm/drmExternals.html
10:15 < pinchartl> Tomohito Esaki from Igel has created a "quick hack" solution and provided a test
10:15 < pinchartl> application. The BSP team would like to know our opinion.
10:15 < pinchartl> morimoto: this may be a stupid question, but
10:15 < pinchartl> the DU doesn't have rendering support, as we have a separate GPU for that
10:16 < pinchartl> why would the DU driver need to create a render node when it can't do rendering ? :-)
10:16 < pinchartl> the GPU node should create the buffers, and they should be imported by the DU driver normally
10:16 < pinchartl> the render node API shouldn't be needed for that
10:16 < morimoto> "DU doesn't have rendering support" = "HW" or "SW " ? I think "SW"
10:17 < pinchartl> neither
10:17 < pinchartl> the DU can compose multiple planes
10:17 < pinchartl> but the DU hardware can't do rendering, it's not a GPU
10:18 < pinchartl> so there's no rendering support in the driver because there's no rendering support in the hardware :-)
10:18 < damm> Can IMR do rendering?
10:18 < morimoto> Ahh, OK, thanks. I was misunderstanding
10:18 < pinchartl> damm: IMR does some kind of rendering
10:19 < pinchartl> as such it would be nice to support it with a DRM driver (and that would create a render node)
10:19 < pinchartl> but it seems that the direction that was chosen was to use V4L2 instead
10:19 < pinchartl> for the GPU I certainly hope V4L2 won't be used :-)
10:19 < damm> i think the input from the m/m group leader is highly regarded
10:19 < pinchartl> but DRM may not either, and that's out of our control
10:20 < pinchartl> I commented on that in the past I believe
10:20 < pinchartl> I think that was when Cogent posted IMR patches
10:20 < pinchartl> (I'd have to check though, it was a long time ago)
10:20 < morimoto> [PATCH v5] media: platform: Renesas IMR driver
10:20 < morimoto> > https://lore.kernel.org/linux-renesas-soc/20170309200818.786255823@cogentembedded.com/
10:20 < morimoto> ?
10:20 < damm> gotcha. thanks
10:22  * kbingham likes that lore is getting good use :)
10:22 < pinchartl> morimoto: that driver, yes. I think I commented on a previous version
10:22 < geertu> Sergei plans to work on v6 soon
10:23 < morimoto> soon. OK thanks. Nice to know
10:23 < morimoto> pinchartl: Thanks. I will feedback this info to BSP team
10:23 < kbingham> https://lore.kernel.org/linux-renesas-soc/1770632.4GFlW6r2cg@avalon/
10:23 < morimoto> s/this/these/
10:25 < morimoto> kbingham: thanks
10:25 < kbingham> pinchartl, https://lore.kernel.org/linux-renesas-soc/15d3888d-e15c-db23-27b4-51d5e2840618@cogentembedded.com/ Konstantin replied saying the name "image renderer is misguiding"
10:25 < pinchartl> it's a bit of a borderline case, depending on whether you think it performs texture rendering, or lens distortion correction
10:26 < geertu> The former is more generic than the latter
10:27 < geertu> Can the latter be implemented on top of the former?
10:27 < pinchartl> possibly, depending on lots of details
10:28 < pinchartl> the use case here, if I understand correctly, is to create the bird-eye-view from four fish-eye cameras
10:28 < pinchartl> so it doesn't just do lens distortion correction
10:28 < pinchartl> it actually creates distortion, just the other way around :-)
10:29 < kbingham> sounds like texture rendering onto a sphere :)
10:30 < wsa_> oh, we do this on c64, too
10:31 < pinchartl> LOL
10:31 < pinchartl> in any case, render nodes shouldn't be needed for the DU as far as I can tell
10:31 < geertu> wsa_: On Amiga, I used the copper for that (only in one dimension, or helix scrolls)
10:31 < geertu> s/or/for/
10:31 < pinchartl> Topic 3. peripericon @FOSDEM 2019
10:32 < pinchartl> the multimedia meeting will be on Tuesday 2019-02-05 from 09:00 to 12:00
10:32 < wsa_> geertu: so, the copper needs render-nodes then? :)
10:33 < geertu> wsa_: No, it doesn't render, it's just a step in the display output pipeline
10:35 < pinchartl> the current agenda is
10:35 < pinchartl> - Up-port status and planning
10:35 < pinchartl> - Planning for the next quarters
10:35 < pinchartl> if you have any other topic to propose, you can submit them here, or reply to the group chat report e-mail later
10:35 < pinchartl> any other topic to discuss for today ?
10:35 < neg> Not from me
10:37 < wsa_> jmondi: please send the preparational material for IO discussion soon
10:37 < jmondi> wsa_: sure, it's on my list for next week
10:37 < jmondi> would you like it before, like tomorrow?
10:37 < wsa_> actually, yes
10:38 < wsa_> I would like at least some days to really prepare
10:38 < jmondi> I actually would then
10:38 < jmondi> s/would/will
10:38 < wsa_> and give other people also time to prepare
10:38 < pinchartl> if there's no other question or topic, I proposed adjourning this meeting. does anyone second ?
10:38 < jmondi> I see... will do
10:38 < jmondi> seconded
10:39 < damm> thanks guys, see ya later
10:39 < wsa_> jmondi: thanks
10:39 < jmondi> wsa_: long overdue
10:40 < pinchartl> meeting adjourned
10:40 < pinchartl> thank you all for attending, and have a great day