path: root/wiki/Chat_log/20170706-mm-chatlog

Multimedia-chat-meeting-2017-07-06

10:04 < pinchartl> Wolfram proposed in an e-mail yesterday to start with multimedia
10:04 < pinchartl> if everybody is fine with that, I'll proceed
10:04 < pinchartl> wsa_: geertu: is that OK ?
10:05 < wsa_> ok
10:05 < geertu> pinchartl: OK
10:05 < pinchartl> let's begin
10:05 < pinchartl> and try to keep it short
10:06 < pinchartl> just as a reminder, we'll start with task status reporting only
10:06 < pinchartl> feel free to ask questions if there's anything you find unclear
10:06 < pinchartl> but please keep the questions limited to task status reporting and planning
10:07 < pinchartl> you can also bring up other topics, which I will then record, but we'll discuss them in the afternoon
10:08 < pinchartl> uli___: I think you were the last to report over e-mail, so you can start
10:08 < uli___> ok
10:08 < uli___> so we found the issue with the silent max9260s, it was an incorrect setup on the blanche board
10:08 < uli___> we can talk to all max9260 now, and the one max9259 that is connected
10:09 < uli___> (there is only one cable)
10:09 < uli___> i have retooled the serdev gpio multiplexing to use the new multiplexer subsystem
10:09 < pinchartl> nice
10:09 < uli___> pleasantly simple, but as yet untested. rsn...
10:10 < uli___> that's it, really
10:10 < uli___> i'll fix this up once it works and send out a new rfc
10:10 < pinchartl> thank you
10:10 < pinchartl> neg: you're next to report from the beach
10:11 < neg> Shorter IRC friendly version. longer one in status update mail, plase us it for the report :-)
10:11 < neg> A)
10:11 < neg>   - Reworked VIN Gen3 DT parsing do deal with not registering the video
10:11 < neg>     device at probe time but instead in the async complete callback().
10:11 < neg>   - Got feedback from Laurent that my idea for incremental async using
10:11 < neg>     subnotifers was not his favorite way to solve this problem, Laurent
10:12 < neg>     also provided feedback on how he thinks is the best way to implement  it.
10:12 < neg>   - Discussed an issue with Kieran about format size reporting in a
10:12 < neg>     media graph pipeline. The issue is for sources providing streams
10:12 < neg>     using the field type TOP, BOTTOM or more common ALTERNATE. Which
10:12 < neg>     format size should they report?
10:12 < neg>   - I'm not working from Stockholm this week but my "secretary" received
10:12 < neg>     a packages from OpenSörce / Renesas yesterday which I'm hoping is
10:12 < neg>     the new Gen3 board and/or the 8-channel camera setup.
10:12 < neg> B)
10:12 < neg>   - Get a working prototype out the door to replace subnotfiers for
10:12 < neg>     incremental async.
10:12 < neg>   - Pickup Kierans ADV7482 DT patches and include them together with
10:12 < neg>     other VIN+CSI-2 DT changes in my for-renesas-drivers branch.
10:12 < neg>   - Hope fully resolve the format size reporting discussions and if
10:12 < neg>     after that there is nothing more I want to solve in my VIN Gen3
10:12 < neg>     branch post it to ML :-)
10:12 < neg>   - Inspect package with new boards and bring them in to my lab and if
10:12 < neg>     it contains a new Gen3 board test how VIN works on it and report to ML.
10:12 < neg> C)
10:12 < neg>   - None
10:12 < neg> Patches of interest)
10:12 < neg>   - [PATCH 0/2] media: v4l: Add support for the Cadence MIPI-CSI2RX
10:12 < neg> --EOT--
10:13 < pinchartl> thank you
10:13 < pinchartl> regarding formats, is that subdev formats ?
10:13 < kbingham> Yes - The subdev format of the ADV748x when using interlaced
10:14 < kbingham> I believe the subdev should report field height (frame_height/2)
10:14 < neg> for me it's about all formats wich can be read from the media graph
10:14 < pinchartl> yes, the height should be halved in that case
10:15 < pinchartl> if the V4L2 documentation isn't clear on this topic, patches are welcome :-)
10:15 < neg> and the SD should report field height and not frame height for ALTERNATE, TOP and BOTTOM formats?
10:15 < kbingham> pinchartl: There is a patch posted on this topic - but I do'nt think it's been integrated yet :)
10:15 < pinchartl> neg: correct
10:16 < neg> and then VIN needs to doubble the height if it uses the HW to interlace the fields before writing it to memory
10:16 < pinchartl> neg: correct as well
10:17 < pinchartl> thank you for reporting patches of interest
10:17 < pinchartl> please all try to keep this in mind
10:17 < neg> OK kbingham thanks for finding the documentaion patch
10:17 < pinchartl> I have that patch series in my inbox, Maxime CC'ed me to get it reviewed -_-'
10:17 < pinchartl> next, kbingham 
10:17 < kbingham> ADV748x: version 6 is posted publicly, and comments from Hans have been addressed to create a v7.
10:17 < kbingham> I hope that this should be acceptable for integration now. The DT patch is being upstreamed though Niklas' work due to inherrent dependancies there.
10:18 < kbingham> The vblank issue is under investigation and prototype patch is posted but a couple of issues to be resolved. Laurent you mentioned you were going to look at this - but did you have time?
10:18 < kbingham> I also anticipate moving to ES2 hardware this week - A board should arrive this week.
10:18 < pinchartl> the board should hopefully arrive today :-)
10:19 < kbingham> And meanwhile I am looking at display list work - although I don't think this is a formal additional task yet :)
10:19 < kbingham> -- eot --
10:20 < pinchartl> regarding the vblank issue, I haven't had time yet (beside fixing the small vblank interrupt issue we discussed yesterday)
10:20 < pinchartl> I can continue, if you prefer taking over, I'm fine with that too, but I remember you had issues reproducing the problem
10:20 < pinchartl> so I propose continuing on my side at least for today
10:20 < kbingham> That sounds good.
10:21 < kbingham> I think the reproduction of the issue was more clear when shutting down kmstest/cube with ctrl-c
10:21 < pinchartl> ok
10:21 < pinchartl> next, jmondi 
10:22 < jmondi> ok!
10:22 < jmondi> A)
10:22 < jmondi> - Sent mx9286 cleanup patches (v1 and v2)
10:22 < jmondi> - Rebased max9286 patches (my series and Laurent's one) on top of Niklas VIN patches v10
10:23 < jmondi> - Implemented skeleton for max9286 subdevice
10:23 < jmondi> - Found out what chip actually is the one I was using to test VIN parallel input
10:23 < jmondi> B)
10:24 < jmondi> - Complete rebase and have capture with a single channel working again
10:24 < jmondi> - Register max9286 subdevice and define DTS layout
10:24 < jmondi> C)
10:25 < jmondi> - Niklas' v10 VIN patch series is based on v4.12: massive DTS re-shuffling and also device tree parsing logic has changed. I now have links in the media controller graph not working
10:25 < jmondi> ^ this can be a topic for afternoon discussions, as I'm not sure what is the best way to re-enable and link VIN nodes with new dts layout
10:25 < jmondi> --eot--
10:26 < pinchartl> thank you
10:26 < pinchartl> the sensor you have identified is an MT9V111, right ?
10:26 < jmondi> yes
10:26 < jmondi> not mt9m111 as I thought it was :/
10:26 < pinchartl> and that's not supported upstream :-/
10:26 < jmondi> nope
10:27 < pinchartl> we'll have to decide how to handle that, as there's a significant amount of work needed to support it
10:27 < dammsan> i'm glad we used migo-r instead of that sensor =)
10:27 < pinchartl> I don't think it's a priority
10:28 < pinchartl> dammsan: to test parallel VIN input ? :-)
10:28 < jmondi> I guess so, you pointed me to another driver for a "similar" chip, but for what I've seen, there's plenty of work to do anyway
10:28 < jmondi> migo-r was for CEU development
10:28 < pinchartl> jmondi: I know, I was teasing Magnus
10:29 < wsa_> jmondi: about MAX9286: a little more detailed "changes since last version" for a patch series would help reviewing quite a bit. otherwise i/we have to find out what you changed in detail.
10:29 < jmondi> that's a pretty nice task: complete CEU driver AND implement mt9v111 sensor driver
10:29 < jmondi> uh, I forgot the v1->v2 changelog?
10:29 < wsa_> jmondi: may I ask you to be a bit more detailed next time?
10:30 < jmondi> yes, not v1->v2 in cover letter, sorry about that
10:30 < pinchartl> jmondi: you mentioned in your e-mail report you will start travelling on Saturday, is it this saturday ?
10:30 < jmondi> this saturday yes
10:30 < jmondi> from next week I'll reduce my working days but I'll be available
10:31 < pinchartl> at what point should Kieran and I take over on the max9286 ?
10:32 < wsa_> jmondi: no big problem :)
10:32 < jmondi> so, this 2 days, I hope I'll be able -at least- to have capture working again with your series and my series rebased on Niklas' v10
10:33 < jmondi> then I have a skeleton implementation of max9286 subdevice to be tested with incremental async registration
10:33 < pinchartl> how about giving you the rest of this week and next week, and then taking over ? would that be good ?
10:33 < jmondi> once that's done, I guess it's time for an handover
10:34 < jmondi> yes, next week I won't work full time but I hope I can complete that and assist with handover
10:34 < jmondi> so it's fine
10:34 < pinchartl> perfect, thank you
10:35 < pinchartl> next: Morimoto-san
10:35 < morimoto> A) What have I done since last time
10:35 < morimoto> I'm continuing ALSA SoC framework cleanup. I got good response from Maintainer about one of its idea.
10:35 < morimoto> But I noticed some issue exist on RFC v1. I posted RFC v2 to ML, no response yet.
10:35 < morimoto> B) What I plan to do till next time
10:35 < morimoto> Continue RFC v2 cleanup for all drivers
10:35 < morimoto> C) Problems I have currently
10:35 < morimoto> No, sir
10:35 < morimoto> --EOT--
10:35 < pinchartl> quick and precise, thank you :-)
10:35 < morimoto> :)
10:36 < pinchartl> are you still supposed not to work on HDMI output for Gen3 due to legal reasons ?
10:36 < morimoto> Renesas Europe(?) guy want to use it. and I can support him
10:36 < morimoto> but, in offically, I still keep distance form it
10:36 < morimoto> on upstream
10:37 < morimoto> maybe not a big problem I think
10:37 < morimoto> but, paper work guys...
10:37 < pinchartl> for now it's fine, I hope it will be solved soon
10:37 < pinchartl> thank you
10:37 < morimoto> Yes, thanks
10:37 < pinchartl> next, dammsan 
10:37 < morimoto> pinchartl: after lunch, I want to confirm about BSP team's request
10:37 < dammsan> i've had time off - nothing new to report
10:38 < pinchartl> morimoto: sure, I haven't forgotten, I noted the questions from your e-mail :-)
10:38 < pinchartl> dammsan: ok, that's an easy one
10:38 < morimoto> thanks
10:38 < pinchartl> are all the Gen3 boards back in the remote farm ?
10:39 < dammsan> good question, i think so =)
10:39 < dammsan> but at new locations
10:39 < dammsan> if there is something special you need then let me know
10:39 < pinchartl> thank you
10:39 < pinchartl> any plan to work on multimedia for the next two weeks ? :-)
10:40 < dammsan> not from my side
10:40 < dammsan> i might refresh ipmmu
10:40 < pinchartl> ok, thank you
10:40 < pinchartl> next, myself
10:41 < pinchartl> - since last meeting
10:41 < pinchartl> posted a new version of the H3 ES2.0 display output patches
10:41 < pinchartl> now that I'm back home all four outputs have been tested
10:41 < pinchartl> (and they work ;-))
10:42 < pinchartl> I've also fixed the DU/VSP initialization order problem
10:42 < pinchartl> but we now have a vblank-related issue when stopping the display
10:42 < pinchartl> for the next two weeks
10:42 < pinchartl> fix the display stop issue
10:43 < pinchartl> clean up the max9286 and rdacm20 driver with Kieran, taking over work form Jacopo during his holidays
10:43 < pinchartl> and hopefully working on the DU automated test suite
10:43 < pinchartl> no issue or blocker for now
10:44 < pinchartl> and for today, I will send the additional tasks descriptions to Magnus
10:44 < pinchartl> sorry for the delay
10:44 < kbingham> pinchartl: Silly question - All four outputs tested ... whats' the fourth ... HDMI0 HDMI1 VGA ...
10:44 < pinchartl> kbingham: LVDS
10:44 < kbingham> pinchartl: Ah of course :)
10:44 < pinchartl> you need a special LVDS panel
10:45 < pinchartl> this closes the task reporting part of the multimedia meeting
10:46 < pinchartl> for this afternoon, I currently have the following topics
10:46 < pinchartl> - question from Morimoto-san
10:46 < pinchartl> - max9286 DT (Jacopo)
10:46 < pinchartl> anything else ?
10:47 < neg> not from me
10:47 < pinchartl> ok, I'll now let Wolfram and Geert decide who goes next :-)
10:47 < pinchartl> thank you all for attending