path: root/wiki/Chat_log/20170201-mm-chatlog

Multimedia-chat-meeting-2017-02-01

09:07 < pinchartl> topics for today
09:07 < pinchartl> - status update
09:07 < geertu> pinchartl: I think Magnus is in a train (with WiFi ;-)
09:07 < pinchartl> - ELC meeting
09:07 < pinchartl> - VIN plans
09:07 < pinchartl> anything else ?
09:07 < neg> not from me
09:08 < pinchartl> (there's also "breakfast at le pain quotidien in brussels" but I think that's just for me today :-))
09:08 < kbingham> pinchartl: We'll add that to tomorrows agenda
09:08 < kbingham> :D
09:09 < pinchartl> :-)
09:09 < pinchartl> Topic 1. Status check for the multimedia tasks
09:09 < pinchartl> Kieran, as you're time constrained, you can start
09:09 < kbingham> Things I've done:
09:09 < kbingham> I've been working on the scaler partition algorithm, and spent some time better understanding the scaler algorithms. I've also worked through the 'pseudo code' from Renesas, and made this compile so I could better understand each of it's actions.
09:09 < kbingham> From this I've added phase calculation into the UDS entity, and I'm trying to rework the algorithm so that each entity "controls it's own fate" and performs it's part of the algorithm. (WPF determines output windows, UDS determines scaling factors, RPF determines source chroma recovery). To achieve needs per partition cropping on each entity - something to talk about F2F later today/tomorrow :D
09:10 < kbingham> Next: FOSDEM, and I'm hoping to get a last minute snowboarding holiday in the school half-term with Keri (13-17th), and more work on PA, and VSP-tests.
09:10 < kbingham> Blockers - only that I want to go through partition algorithm design with Laurent, which is conveniently possible in the next couple of days.
09:10 < kbingham> (pre-written due to time constraints :D )
09:11 < kbingham> EOT
09:11 -!- morimoto [~user@relprex1.renesas.com] has joined #periperi
09:11 < pinchartl> thank you
09:11 < pinchartl> hello Morimoto-san
09:11 < morimoto> Hi
09:11 < pinchartl> we have just started with Kieran's update
09:11 < morimoto> Oops ? already started ??
09:12 < kbingham> morimoto: my fault-  I have a plane to catch :D
09:12 < pinchartl> has Japan moved its time zone by 15 minutes without any notice ? :-)
09:13 < morimoto> I'm booking today's meeting started from 17:30, now Japan is 17:10
09:13 < pinchartl> I think I've sent the invitation starting from 17:00
09:13 < pinchartl> but no worries, you're here :-)
09:13 < pinchartl> I'll go next to give you a bit of time to prepare
09:13 < morimoto> Grr, sorry about that
09:13 < pinchartl> since last meeting
09:14 < pinchartl> I've continued advancing the HDMI output driver towards merging
09:14 < pinchartl> there's few blockers left
09:14 < pinchartl> Neil Armstrong from BayLibre has submitted a patch series to prepare the dw-hdmi driver to be used in Amlogic SoCs. Discussions on how to integrate both patch series are complete.
09:14 < pinchartl> but it means a bit more work for me as they need to decouple PHy handling completely
09:14 < pinchartl> (because they use an external PHY)
09:15 < pinchartl> I mean a vendor PHY
09:15 < pinchartl> I've also had a meeting in Oslo with Hans Verkuil and Sakari Ailus about V4L2 and media controller race conditions
09:16 < pinchartl> with a whole day preparation meeting in Helsinki with Sakari the day before
09:16 < pinchartl> the meetings went well, we agreed on the way forward, and Sakari has posted a summary to the linux-media mailing list
09:16 < pinchartl> (http://www.spinics.net/lists/linux-media/msg110310.html)
09:16 < pinchartl> the next step is to convince Mauro
09:16 < pinchartl> it looks pretty good so far
09:17 < pinchartl> he has pushed back a bit oon a few points
09:17 < pinchartl> but nothing too bad
09:17 < pinchartl> and Hans is on our side
09:17 < pinchartl> and finally I had holidays last week
09:17 < morimoto> Enjoyed sky ?
09:17 < pinchartl> very much :-)
09:18 < morimoto> good :)
09:18 < pinchartl> oh, I also had a long discussion about how to move forward with the V4L2 metadata API, needed for histogram support in the VSP driver
09:18 < pinchartl> for the next two weeks
09:18 < pinchartl> - FOSDEM
09:18 < pinchartl> - ELC preparation
09:19 < pinchartl> - Gen3 HDMI output
09:19 < pinchartl> - patch review
09:19 < pinchartl> no issue or blocker
09:19 < pinchartl> that's it for me
09:19 < pinchartl> neg: can you go next ?
09:19 < neg> sure
09:19 < neg> A) Posted a few VIN and CSI-2 series (see status mail for list) and partaken in discussions about adding multiple subchannels for VIN/CSI-2 on Gen3
09:20 < neg> B) Attend FOSDEM and talk about VIN issues with Laurent and monitor feed back to patches posted in A as one of the blocks work for IO group
09:20 < neg> C) Wishihg to get feedback on how to speed up VIN/CSI-2 patch acceptence rate :-)
09:20 < neg> EOT
09:21 < pinchartl> morimoto: VIN is a topic of this meeting, we'll get to it right after you
09:21 < pinchartl> neg: thank you
09:21 < morimoto> pinchartl: thanks
09:22 < pinchartl> morimoto: your turn
09:22 < morimoto> A. Since last time
09:22 < morimoto> I talked with Rob (= OF-graph maintainer) and ML people about ALSA SoC OF-graph DT bindings, and almost OK I believe.
09:22 < morimoto> I posted OF-graph ALSA SoC support patch-set this week.
09:22 < morimoto> I noticed we can solve HDMI sound port detection issue of OF-graph if we have new ALSA SoC function (before I produced "type=sound" property for it. Laurent indicated media-controller for it). I implemented it, and it works.
09:22 < morimoto> I got few BSP team request for you guys ;)
09:22 < morimoto> B. Until next time
09:23 < morimoto> I will post HDMI sound patch-set if OF-graph patch-set was accepted.
09:23 < morimoto> And paper work for MAX Camera
09:23 < morimoto> C. Blockers
09:23 < morimoto> There is no response from Rob
09:23 < morimoto> CSI2 camera export needs Happy lovely paper work
09:23 < morimoto> --EOF--
09:23 < morimoto> About MAX Camera, BSP team want to have this support until 3/E, maybe this is not realistic ?
09:23 < morimoto> Export paper work takes more time, and no develop time for you guys :(
09:23 < morimoto> And this MAX Camera will be rental, we need to back it to BSP team.
09:23 < morimoto> So, my plan is we will buy it.
09:24 < morimoto> But, it seems buying will takes time
09:24 < morimoto> some plan we have
09:24 < morimoto> 1)
09:24 < morimoto> send rental camera to you guys
09:25 < morimoto> and Renesas buy it in parallel
09:25 < morimoto> And someday, we will switch
09:25 < morimoto> or
09:25 < morimoto> 2)
09:25 < morimoto> I will ask you guys to buy it from Europa
09:25 < pinchartl> thank youcan it be bought easily ? do you know what the lead time is ?
09:25 < morimoto> but, 2) is not good for time limit
09:26 < morimoto> Acorrding to HW guys, if you buy it today, you can get it 3/E
09:26 < pinchartl> :-)
09:26 < morimoto> Now we are considering about it
09:26 < morimoto> Ahh
09:27 < pinchartl> I'm fine with that, but I suppose the BSP team iisn't :-)
09:27 < morimoto> about MAX chip datasheet, it is under NDA
09:27 < morimoto> We are solving export paper work
09:28 < pinchartl> can we get the MAX9286 datasheet under NDA ?
09:28 < morimoto> pinchartl: if we can avoid BSP team's time limit, can I ask you buy it from Europe ?
09:28 < morimoto> We will get more detail. URL, cost, serial number etc...
09:29 < pinchartl> morimoto: sure, assuming that Maxim will sell it to us (as the documentation is under NDA there could also be restrictions on buying the hardware)
09:29 < morimoto> pinchartl: about MAX datasheet, yes, we can get it nunder NDA. we need more detail infromation from HW guys.
09:30 < morimoto> It needs additional board + camera. camera is very generic. no problem
09:30 < pinchartl> if you tell us what to buy we can certainly buy it, no problem with that
09:31 < morimoto> but, board is R-Car special. We need special interface guy (?)
09:31 < morimoto> pinchartl: yes, thanks
09:31 < morimoto> We are now planing how to solve this issue.
09:31 < morimoto> thats it from me
09:32 < pinchartl> the interface board is a custom PCB from Renesas, right ?
09:32 < morimoto> maybe, I don't know detail yet
09:33 < pinchartl> I would assume so
09:33 < pinchartl> we will need that PCB
09:33 < pinchartl> when do you expect to have information on what is needed ?
09:34 < morimoto> Now, HW guys is collecting information
09:34 < morimoto> BSP team -> HW guys -> ???
09:34 < morimoto> I don't know how long
09:35 < pinchartl> ok
09:35 < pinchartl> we *could* try to get different hardware to work on
09:35 < pinchartl> but I think it will take longer
09:35 < pinchartl> as we would likely have to develop an interface PCB anyway
09:35 < morimoto> I guess is this related to 2nd Q1 SoW ?
09:36 < neg> If there is a upstream driver for a CSI-2 transmitter that supports multiple channels that might still be less time to develop a interface PCB
09:36 < morimoto> If so, when do you need it for Q1 2nd SoW ?
09:37 < morimoto> I guess, we can collect many information this and next week (?)
09:37 < morimoto> And start export paper work +1 or +2 weeks
09:38 < morimoto> And ship it to Europe, it take 1 week ?
09:38 < morimoto> this is *BEST* case
09:38 < pinchartl> this brings us to the next topic, VIN/CSI-2 planning :-)
09:38 < pinchartl> I don't think 3/E is a reasonable target
09:38 < morimoto> Hmm.. yes
09:38 < pinchartl> as Niklas explained in his last e-mail, there is lots of work needed
09:38 < morimoto> Yes, yes, I know
09:38 < pinchartl> we could possibly produce a prototype by 3/E
09:38 < pinchartl> *if* we can assign enough developers on this
09:39 < morimoto> without HW ?
09:39 < pinchartl> Niklas (unless that has changed) works 50% for Renesas
09:39 < pinchartl> (25% for the base contract and 25% for the additional tasks)
09:39 < pinchartl> that's now enough time until 3/E to implement a solution
09:39 < pinchartl> no, with hardware
09:39 < morimoto> OK
09:39 < pinchartl> hardware isn't needed at the very beginning
09:39 < pinchartl> but we'll need it for testing
09:40 < morimoto> and remote machine is not enough ?
09:40 < pinchartl> it depends on the hardware, whether it can be controlled remotely
09:40 < pinchartl> I don't expect things to work out of the box
09:40 < pinchartl> remote access has a high latency
09:40 < morimoto> OK
09:41 < pinchartl> 5 minutes to boot a new kernel
09:41 < morimoto> Ahh, not good
09:41 < pinchartl> as we'll need lots of tests to debug the code, that would slow things up
09:41 < morimoto> I will do my best for you guy ;)
09:41 < pinchartl> I mean slow things down :-)
09:41 < pinchartl> thank you
09:41 < pinchartl> another thing
09:41 < neg> thank you :-)
09:42 < pinchartl> as I said, producing a prototype for 3/E requires lots of work
09:42 < pinchartl> Niklas won't have enough time alone
09:42 < pinchartl> so we'll need to work together
09:42 < pinchartl> Kieran and I can help
09:42 < pinchartl> but then we won't have enough time to solve the VSP/DU flicker issues
09:42 < morimoto> I think BSP team's planning is already broken. So I can tell it to them
09:43 < pinchartl> they have to pick the priority
09:43 < pinchartl> the second half of Q2 is one month
09:43 < pinchartl> 20 days
09:43 < pinchartl> as Niklas is assigned to Renesas at 50%, that's 10 days for him
09:43 < pinchartl> 5 days of base task, 5 days of additional task
09:44 < pinchartl> it makes no sense to split that between VIN and something else
09:44 < pinchartl> so if we need to focus on VIN, I'd like Niklas to spend all his time on VIN in the second batch of Q1, and in Q2
09:44 < pinchartl> otherwise we won't move forward if we split this in 5 days tasks here and there
09:45 < pinchartl> another thing
09:45 < pinchartl> if we assign enough developers to this 3/E might be possible for a prototype
09:45 < pinchartl> *but*
09:45 < pinchartl> it really means 3/E
09:45 < pinchartl> we can't do it for 3/M
09:45 < pinchartl> which is when Jinzai expects reports
09:46 < pinchartl> in that case I'll want to send reports at 3/E
09:46 < pinchartl> as spending time on reports for 3/M while we rush to finish the code for 3/E makes absolutely no sense
09:47 < morimoto> I guess we can avoid 3/E limitation for VIN for now. And I think I need to talk to my Boss about Niklas time ?
09:47 < pinchartl> and to Niklas and Magnus too I assume :-)
09:47 < morimoto> Ahh yes. ELC is good time for it ?
09:47 < morimoto> Oops, too late ?
09:47 < pinchartl> it depends on the priorities :-)
09:48 < pinchartl> if the BSP team thinks this is the highest priority and we need to start working on it immediately, then ELC is late
09:48 < pinchartl> if we can have a delay, it's fine
09:49 < pinchartl> my point is that, if Niklas has 15 days of base task and and 15 days or additional tasks per quarter, it makes no sense to split that between multiple tasks for multiple groups if VIN is the focus
09:49 < pinchartl> otherwise the task scheduler overhead will be too high :-)
09:49 < pinchartl> neg: please feel free to chime in as we're talking about you :-)
09:49 < neg> :-)
09:49 < morimoto> BSP team 3/E request today is Fence, DU init order fixup, suspend/resume bug fix
09:50 < morimoto> VIN was requested, but it is too late
09:50 < pinchartl> given the amount of work for VIN, splitting the work in 5 or 10 days tasks spread over 6 months doesn't make sense to me. it should be grouped
09:50 < pinchartl> morimoto: I think that's a more reasonable plan
09:50 < neg> I agree with what you are saygin, I can't do this by my self we need to coalaborate and it make sens for me to focus mor then 1/3 of my time on this as I do now for multimedia tasks
09:51 < morimoto> OK, I hope we can talk this F2F on ELC. I will correct information etc for it
09:52 < morimoto> BSP plan, budgetcl
09:52 < neg> pinchartl: do you know if Sakari will be at FOSDEM or ELC?
09:52 < pinchartl> sure, we'll certainly talk about it at ELC
09:52 < pinchartl> neg: he will unfortunately not attend either :-(
09:53 < pinchartl> neg: how about spending a day or two in Helsinki ?
09:53 < neg> Hum maybe it would make sens for me to fly to Helsinki and meet both of you and talk F2F about his current work on subchannels and if we can colaborate forward?
09:53 < pinchartl> :-)
09:54 < pinchartl> morimoto: do you think we could get budget for Niklas to spend two days in Helsinki ? Stockholm-Helsinki is a very short flight, it will be cheap
09:54 < morimoto> for FOSDEM ? or meeting ?
09:54 < pinchartl> for meeting with me and Sakari Ialus in Helsinki
09:54 < pinchartl> Sakari works for Intel
09:55 < morimoto> OK, please wait
09:55 < pinchartl> and has started working on a V4L2 API for multiple virtual channels
09:55 < pinchartl> if we can spend one or two days with him face to face it would help a lot
09:55 < pinchartl> but he won't attend the FOSDEM or ELC
09:55 < pinchartl> so Niklas could fly to Helsinki, that's very close
09:56 < pinchartl> and Sakari can likely host a meeting at the Intel office there
09:56 < morimoto> pinchartl: neg: we can get OK from our boss v(^-^)v
09:56 < pinchartl> we're talking about a few hundred euros to save weeks of e-mail discussions, I think it's worth it :-)
09:57 < pinchartl> that will be after the FOSDEM anyway, so we have a bit of time
09:58 < morimoto> FOSDEM -> Helsinki -> ELC. business-man !!
09:58 < neg> It's a hard life :-)
09:59 < pinchartl> neg: seriously, spending all your time in planes *is* a hard life
09:59 < pinchartl> I think that's it for the VIN discussion for today, we'll talk more about it face to face in Brussels
09:59 < pinchartl> and this brings us to the next topic, FOSDEM preparation
09:59 < pinchartl> neg: you will arrive on Thursday evening, right ?
09:59 < neg> yes
10:00 < pinchartl> I wonder when we should schedule time for face to face dicussions
10:00 < pinchartl> Friday will be busy with core and I/O
10:00 < pinchartl> maybe Thursday evening ?
10:00 < neg> arrive 17:20 and hope I can make it to my hotell by 18:30 I think it should be doable
10:01 < pinchartl> we can either have a late dinner, or talk during/after dinner
10:01 < neg> so any time between that and sunday evening works for me except ofc for periperi meetings during Friday
10:02 < neg> Thursday evening works, maybe it's good to have it before/after dinner so I can make notes
10:02 < pinchartl> how about having dinner at 19:00 and discuss after that ?
10:02 < pinchartl> we can talk about planning and other less technical topics during dinner
10:02 < neg> yes sounds like a good plan
10:02 < pinchartl> and move the technical parts after dinner
10:03 < pinchartl> OK, I'll book a table for 19:00 then
10:03 < pinchartl> for 5 of us
10:03 < neg> nice, at what location?
10:04 < pinchartl> I don't know yet, I'll sort it out today
10:04 < neg> ahh nice, thanks for arraigning it
10:05 < pinchartl> you're welcome
10:05 < pinchartl> next topics, ELC
10:05 < pinchartl> there will be no V4L2 meeting at ELC
10:06 < pinchartl> so Monday and Friday are free
10:06 < pinchartl> I this propose having a multimedia meeting on Monday
10:06 < morimoto> Nice for me
10:06 < neg> works for me
10:07 < pinchartl> I assume it will owrk for Magnus too
10:07 < morimoto> I and Magnus will arrive there at 17th
10:07 < pinchartl> Ulrich and Kieran don't plan to attend ELC as far as I know
10:07 < pinchartl> but it could be possible to change their minds
10:07 < pinchartl> (assuming we can get budget)
10:08 < morimoto> about budget, maybe no problem.
10:10 < pinchartl> :-)
10:10 < pinchartl> last topic then, next meeting
10:10 < pinchartl> I propose two weeks from now
10:10 < pinchartl> February the 15th
10:10 < pinchartl> same time as today (that's 17:00 JST ;-))
10:10 < morimoto> Oops, Feb 15th doesn't work for me
10:11 < pinchartl> how about the 14th or the 16th ?
10:11 < morimoto> both 14th, 16th are OK
10:11 < neg> both works for me
10:11 < pinchartl> when is the next core group chat planned ?
10:12 < geertu> pinchartl: Feb 14th?
10:12 < geertu> Feb21 is ELC
10:13 < geertu> Or we can skip directly to Feb 28 :-)
10:13 < morimoto> 28 is OK for me :)
10:13 < pinchartl> it doesn't make much sense to have a multimedia meeting on the 16th and another one at ELC on the 20th
10:13 < pinchartl> should we skip to the 20th directly ?
10:14 < neg> works for me, how do that allign for kbingham and uli time zone wise?
10:14 < morimoto> Do you mean skip 20th meeting on America ?
10:14 < pinchartl> morimoto: yes
10:15 < morimoto> Oops, it is not good for me. I need to report about it :(
10:15 < geertu> "skip to" != "skip"
10:15 < pinchartl> morimoto: no sorry
10:15 < pinchartl> I meant skipping the meeting on the 16th
10:15 < geertu> neg: same time difference as JP vs EU
10:15 < morimoto> Ah, next meeting *is* 20th ?
10:15 < pinchartl> skipping to the 20th means skipping the 16th and going directly to the 20th
10:16 < pinchartl> there will be one face to face meeting on the 20th
10:16 < morimoto> OK, thanks. it is very good for me :)
10:16 < pinchartl> so I propose skipping the one on the 16th, it makes little sense to have two meetings just a few days apart
10:16 < pinchartl> let's do that then
10:17 < pinchartl> morimoto: you will arrive on the 17th, right ?
10:17 < morimoto> Yes
10:17 < pinchartl> neg: how about you ?
10:17 < pinchartl> morimoto: and when will you leave ?
10:17 < neg> 19th
10:17 < morimoto> 24th
10:17 < morimoto> with Magnus
10:18 < pinchartl> neg: same question, when will you leave ?
10:18 < neg> 25th :-)
10:18 < pinchartl> morimoto: does Magnus arrive on the 17th too ?
10:18 < morimoto> Yes. same plane
10:18 < pinchartl> :-)
10:19 < pinchartl> I think that's it for today
10:19 < pinchartl> I propose adjourning the meeting
10:19 < pinchartl> does anyone second ?
10:19 < neg> seconded
10:19 < morimoto> 2nd !
10:20 < pinchartl> meeting adjourned
10:20 < pinchartl> thank you all for attending