path: root/wiki/Chat_log/20171214-mm-chatlog

Multimedia-chat-meeting-2017-12-14

10:10 < pinchartl> looks like we have a full hosue today
10:10 < pinchartl> hello everybody
10:10 < pinchartl> let's get started with status updates
10:10  * kbingham[m] should move to a real keyboard
10:11 < pinchartl> I'll copy & paste from e-mail to make it faster
10:11 < pinchartl> first, uli___ 
10:12 < pinchartl> Since last meeting:
10:12 < pinchartl> - Forward-ported Rogue GPU driver on M3-W to mainline.
10:12 < pinchartl> OpenCL and proprietary API tests pass 100% on the mainline M3-W GPU driver.
10:12 < pinchartl> OpenGL tests (both shipped with the binary blobs and home-brewed tests) fail early, though. eglInitialize() not very helpfully returns EGL_NOT_INITIALIZED.
10:12 < pinchartl> This is suspected to be a user space problem, will try to find a user space/binary blob combination that works.
10:12 < pinchartl> For the next two weeks:
10:12 < pinchartl> - Mix and match the various versions of GPU user space libraries in
10:12 < pinchartl> an effort to get OpenGL to work
10:12 < pinchartl> Issues and Blockers: None
10:12 < pinchartl> uli___: anything to add?
10:12 < uli___> that's it, really
10:12 < pinchartl> thank you
10:12 < pinchartl> next, Kieran
10:12 < pinchartl> Since last meeting:
10:12 < pinchartl> - Rebased GMSL/base on top of renesas-drivers-v4.15-rc1 with Niklas' help
10:12 < pinchartl> - Eagle-V3M GMSL support
10:12 < pinchartl> - Implemented a VIN loopback test
10:12 < pinchartl> Frames are compared, and invalid frames stored for post-analysis
10:12 < pinchartl> - HDMI Interlaced support query
10:12 < pinchartl> For the next two weeks:
10:12 < pinchartl> - Christmas Holidays
10:12 < pinchartl> Issues and Blockers: None
10:12 < pinchartl> kbingham[m]: anything to add ?
10:13 < pinchartl> (or to fix, as I have reformatted the text from e-mail status updates for the report)
10:13 < pinchartl> or kbingham maybe ?
10:13 < neg> kbingham[m]: good work with the looptests!
10:13 < kbingham> pinchartl: Not since I put those in an email at 1am no :)
10:13 < pinchartl> :-)
10:14 < pinchartl> next, Jacopo
10:14 < pinchartl> Since last meeting:
10:14 < pinchartl> - Rebased and retested existing GMSL code
10:14 < pinchartl> - Tested and compared latest Cogent's code drop
10:14 < pinchartl> About aone month ago we have received a new code drop from Cogent (https://
10:14 < pinchartl> github.com/CogentEmbedded/meta-rcar/blob/v2.23.0/meta-rcar-gen3-adas/recipes-
10:14 < pinchartl> kernel/linux/linux-renesas/0030-Gen3-LVDS-cameras.patch). It differs from our current code in OV10635 settings and MAX9286/MAX9271 settings. Using Cogent's settings has no impact on frame synchronization.
10:14 < pinchartl> - Added debug output to v4l2-async
10:14 < pinchartl> - Started building Yocto for GMSL setup to test Cogent's code.
10:14 < pinchartl> Renesas BSP layer built, still have to build Cogent's layer.
10:14 < pinchartl> For the next two weeks:
10:14 < pinchartl> - Collect CEU feedbacks and submit v2
10:14 < pinchartl> - OV7670 device tree parsing
10:14 < pinchartl> - Update gmsl/base with kieran's new tag and test image capture on v4.15-rc1
10:14 < pinchartl> - Yocto build of Cogent's BSP
10:14 < pinchartl> Issues and Blockers:
10:14 < pinchartl> - With v4.15-rc1 GMSL devices are registered but no frame could be captured
10:14 < pinchartl> - GMSL frame sync still no working and becoming a blocker
10:14 < pinchartl> jmondi: anything to add/fix ?
10:14 < kbingham> neg: Yes, it's certainly an interesting way forwards once we get through a few minor format issues :)
10:15 < jmondi> just wanted to know from neg if it expected to reecive an -EPIPE when capturing on v4.15-rc1
10:15 < neg> jmondi: then the pipeline verification fails
10:15 < jmondi> (ie. because of missing mux support)
10:16 < neg> jmondi: we can take this after the meeting and try to bang it out? I'm also porting the new mux setup to the GMSL branch so will probobly run in to the same issues you have
10:16 < jmondi> and I should resend my last series to have gmsl device node created to better exaplain why we need that on v4.15-rc1
10:17 < jmondi> neg: sure
10:17 < pinchartl> in that case
10:17 < pinchartl> next, Niklas
10:17 < neg> jmondi: The old mux implementation should be there in the GMSL branch but I have not tested it ontop of v4.15-rc1, only the new one and it wroks for adv748x at least
10:17 < pinchartl> Since last meeting:
10:17 < pinchartl> - Added VIN support to V3M
10:17 < pinchartl> - [PATCH] rcar-vin: enable support for r8a77970
10:17 < pinchartl> - [PATCH 0/2] rcar-csi2: enable support for r8a77970
10:17 < pinchartl> - [PATCH 0/3] arm64: dts: renesas: r8a77970: connect VIN to CSI40
10:17 < pinchartl> - [PATCH v12 0/2] rcar-csi2: add Renesas R-Car MIPI CSI-2
10:17 < pinchartl> - [PATCH v8 00/28] rcar-vin: Add Gen3 with media controller
10:17 < pinchartl> - [PATCH v9 00/28] rcar-vin: Add Gen3 with media controller
10:17 < pinchartl> - Multiplexed pads based on routing and framedesc patches from Sakari and Laurent done for rcar-vin, rcar-csi2 and adv748x
10:17 < pinchartl> - User-space patches for v4l2-ctl to support routing operations needed to control multiplexed pads from user-space
10:17 < pinchartl> - Begun adopting 8-camera patches for new multiplexed pads
10:17 < pinchartl> - Confirmed CEC functionality on Koelsch (for Lager test-case). Found problems with CEC and adv7612. Hans have hack patch to make it works but it's not ready to be upstream but works to loopback test CEC on Gen2.
10:17 < pinchartl> - Started to address Laurent's review comments to rcar-vin Gen3
10:17 < pinchartl> For the next two weeks:
10:17 < pinchartl> - Post multiplexed streams kernel and user-space patches
10:17 < pinchartl> - If time try to help Hans by adding and testing CEC on Lager, hopefully with the help from Laurent who kindly offered to test on his Lager.
10:17 < pinchartl> Issues and Blockers: None
10:17 < pinchartl> neg: anything to add/fix ?
10:18 < neg> pinchartl: only that CEC test on Lager will be postponed to next year :-)
10:18 < pinchartl> indeed :-)
10:18 < pinchartl> sorry about that, I thought I would have time today but unfortunately I won't
10:18 < neg> no problems
10:19 < pinchartl> well, it could still be done before the next meeting
10:19 < pinchartl> next, Morimoto-san
10:20 < pinchartl> - ALSA SoC cleanup prepare patches have been accepted
10:20 < pinchartl> Main cleanup patches are very big, still trying to find out when the best time to post them would be.
10:20 < pinchartl> - Fixes for sound issues reported by the BSP team
10:20 < pinchartl> - M3N Salvator-XS board export paper work
10:20 < pinchartl> Until next meeting:
10:20 < pinchartl> - Post main ALSA SoC cleanup patches
10:20 < pinchartl> - M3N Salvator-XS board export paper work
10:20 < pinchartl> Issues and Blockers: None
10:20 < geertu> dammsan's remote lager with manual fiddling on his side?
10:20 < pinchartl> anything to add/fix ?
10:20 < pinchartl> morimoto`: anything to add/fix ?
10:20 < morimoto`> nothing,thanks
10:21 < pinchartl> you're welcome
10:21 < pinchartl> next, me
10:21 < pinchartl> Since last meeting:
10:21 < pinchartl> - More patch review
10:21 < pinchartl> - Submitted DISCOM support to calculate CRC in the VSP
10:21 < pinchartl> - Nearly finished display color keying support for Gen3
10:21 < pinchartl> - Submitted new version of the DU plane boundaries fixes
10:21 < pinchartl> - Tested DU dmabuf import fixes locally
10:21 < pinchartl> Until next meeting:
10:21 < pinchartl> - Submit display color keying support for Gen3
10:21 < pinchartl> - More patch review
10:21 < pinchartl> - New version of the V4L2 unbind/userspace race condition handling
10:21 < pinchartl> - Start reworking video device registration code
10:21 < pinchartl> Issues and blockers: None
10:21 < pinchartl> and I agree with myself
10:21 < pinchartl> dammsan: anything to report on your side ?
10:22 < pinchartl> geertu: have we agreed on whether the next meeting will be on the 9th or the 11th ?
10:24 < pinchartl> while waiting for answers, let's move to the questions from Morimoto-san
10:24 < morimoto`> pinchartl: dammsan said "no"
10:25 < pinchartl> - Color Key support for DU/VSP
10:25 < pinchartl> This is due for tomorrow, I will post patches today
10:25 < morimoto`> thanks!
10:25 < pinchartl> - Calculate DU DPLLCR to have smaller jitter
10:25 < pinchartl> I've replied to your e-mail, I assume you will send a v3
10:26 < morimoto`> Yes, thanks.
10:26 < pinchartl> - Interlace height calculation mismatch between HDMI <-> CVBS
10:26 < pinchartl> kbingham: ?
10:26 < pinchartl> - Remove "VNMC_REG = 0" writing from VIN driver
10:26 < pinchartl> neg: ?
10:27 < kbingham> - Ah - yes - I need to look at the interlace issue and send a patch I believe.
10:27 < neg> pinchartl: "VNMC_REG = 0" is addressed in the later vin series
10:27 < pinchartl> neg: thank you
10:27 < pinchartl> kbingham: when do you plan to do so ?
10:27 < morimoto`> neg: upsteamed already, do you mean ?
10:28 < kbingham> pinchartl: I find myself with very little spare time between now and Christmas ... If this is a high priority task - then I'll make time - but otherwise this will be a new year task for me.
10:28 < neg> morimoto`: no, it's part of VIN Gen3 patches which is not yet upstream
10:29 < kbingham> It's probably not a long task ...
10:29 < morimoto`> kbingham: no problem. it is not a high priority. I just asked
10:29 < pinchartl> kbingham: can we say for the next meeting then ? (9th or 11th, pending geertu's answer)
10:29 < kbingham> morimoto`: Ok thanks -
10:29 < neg> morimoto`: but the fault is also introduced by erlier versions of the same patch-set, so the fault have never been present upstream
10:29 < kbingham> pinchartl: Yes, I would hope January is achieveable :)
10:29 < pinchartl> thank you
10:30 < morimoto`> neg: OK, sounds nice. Thanks
10:30 < pinchartl> - ADV7511W slave address conflict issue on D3
10:30 < pinchartl> I haven't had time to work on this yet, I will address it as part of my base contract, planned for the first two weeks of January as well
10:31 < morimoto`> pinchartl: thanks. no stress
10:32 < pinchartl> - drmModeAtomicCommit() API test makes VGA side bad
10:32 < pinchartl> I have confirmed the issue but not investigated it yet
10:32 < pinchartl> I'm not sure yet how big the problem is and whether an additional task would be needed
10:32 < pinchartl> how urgent is it ?
10:32 < morimoto`> Hmm.. not sure.
10:33 < pinchartl> it's pretty bad, so I'd like to prioritize it
10:33 < morimoto`> it can be next add. task ?
10:33 < pinchartl> yes, I think it's a good idea
10:33 < morimoto`> OK, thanks
10:33 < pinchartl> I think those were all your questions
10:34 < pinchartl> next topic, FOSDEM
10:35 < pinchartl> easy question first, who wants to join the multimedia dinner on Thursday evening ?
10:35 < jmondi> o/
10:35 < neg> me
10:35 < morimoto`> pinchartl: I confirmed about drmModeAtomicCommit() priority. It is HI priority
10:36 < pinchartl> morimoto`: ok, thank you
10:36 < morimoto`> pinchartl: +1 topic is DU separation.
10:36 < pinchartl> kbingham: for virtualization ?
10:36 < pinchartl> I meant
10:36 < morimoto`> pinchartl: this is low priority
10:36 < pinchartl> morimoto`: for virtualization ?
10:36 < morimoto`> Yes
10:36 < pinchartl> ok
10:37 < pinchartl> let's discuss it as part of the additional tasks
10:37 < morimoto`> OK, thanks.
10:37 < pinchartl> uli___: no plan to attend the multimedia meeting on Thursday before the FOSDEM ?
10:37 < uli___> no definite plan for now
10:38 < pinchartl> ok
10:39 < pinchartl> ok
10:39 < pinchartl> then
10:39 < pinchartl> it seems our plan to organize a code camp during the week after the FOSDEM in a remote location that would force us to work needs to be cancelled
10:39 < pinchartl> be clearly can't do so without at least one Japanese attendee
10:40 < pinchartl> I would, however, still like to organize a GMSL code camp
10:40 < pinchartl> we've discussed this privately before, the idea would be to work together in Brussels before the FOSDEM from Monday to Thursday
10:40 < pinchartl> Jacopo, Kieran and Niklas have expressed their interest
10:41 < pinchartl> dammsan: morimoto`: I assume you won't be able to join
10:41 < pinchartl> but would this be feasible from a budget point of view ?
10:41 < pinchartl> assuming we can extend the apartment booking we have for Thursday-Monday, this will be cheap
10:42 < pinchartl> as the four of us can stay in the same apartment and work from there
10:43 < pinchartl> kbingham: how much would be the extra cost ?
10:43 < pinchartl> for 4 additional nights
10:43 < kbingham> I'll check the booking
10:44 < morimoto`> pinchartl: unfortunately, we can't join
10:45 < pinchartl> morimoto`: I wonder when our next face-to-face meeting should be. I will not attend ELC this year
10:45 < kbingham> I believe the apartment is €180 per night.
10:45 < pinchartl> so that's 45€ / person
10:46 < morimoto`> pinchartl: Good question, and difficult to say
10:46 < morimoto`> It seems we can't goto ELC too, at this point. but not sure
10:47 < pinchartl> do you have more information about those budget cut ? does it affect other things than travel ? how long is it expected to last ?
10:47 < jmondi> morimoto`: what's happening? are they trapping in you in Japan this year?
10:48 < morimoto`> In Renesas side, our group were changed. Then, budget was changed too.
10:48 < morimoto`> I don't know detail, but, it seems some kind of confusable things happen
10:49 < morimoto`> I can say this is paper work issue
10:49 < pinchartl> so there's no need for all of us to start looking for a new job today ? :-)
10:49 < morimoto`> pinchartl: no problem: )
10:49 < pinchartl> :-)
10:50 < morimoto`> EuroPeri member are very important for us
10:50 < pinchartl> JapanesePeri is very important for us too :-)
10:50 < neg> Ohh and I already sent out my resume to McDonalds ;-)
10:50 < kbingham> neg: You're not qualified to work at McDonalds. .. you don't have the required skill set :)
10:51 < neg> kbingham: I know, I was hoping for som on-site traning on burger flipping
10:51 < jmondi> morimoto`: new fight with paperwork guy ahead?
10:51 < morimoto`> But because of this budget paper work, Renesas member (= morimoto, shimoda) will have issue next year
10:51 < pinchartl> anyway, I think we can move forward with our plan to work on GMSL from Monday to Thursday
10:51 < pinchartl> morimoto`: not a big issue I hope
10:52 < morimoto`> jmondi: not myside, but budget paper work guys
10:53 < pinchartl> jmondi, kbingham, neg: please plan to start working on Monday morning when you make your travel plans for the FOSDEM
10:53 < pinchartl> I will personally arrive on Sunday evening
10:53 < jmondi> morimoto`: budget paperwork guy is some kind of eveolution of regulare paperwork guy, sort of next level enemy?
10:53 < jmondi> we should get there on Sunday then
10:53 < neg> pinchartl: I also plan to arrive on Sunday
10:54 < jmondi> 28th then!
10:55 < morimoto`> jmondi: they have very hard stone head...
10:55 < pinchartl> next topic, additional tasks for 2018 Q1
10:56 < pinchartl> are there tasks you would like to work on in particular ?
10:56 < pinchartl> I have a list, let me paste it
10:57 < pinchartl> * GMSL
10:57 < pinchartl> - Camera instantiation through DT overlay
10:57 < pinchartl> - Camera auto-detection from list of known types
10:57 < pinchartl> - RDACM21
10:57 < pinchartl> - GMSL capture and display test using KMSxx
10:57 < pinchartl> - More GMSL work...
10:57 < pinchartl> (there's more)
10:57 < shimoda> sorry, i'll go home now. see you next time!
10:58 < pinchartl> * Testing
10:58 < pinchartl> - Run igt on the DU
10:58 < pinchartl> - Integrate DISCOM in DU test suite
10:58 < pinchartl> - JTAG Support for Gen3 platforms (more of a Core task)
10:58 < pinchartl> - Python helpers for media-controller
10:58 < pinchartl> - More work on VIN loopback
10:58 < pinchartl> * New hardware
10:58 < pinchartl> - Display Output Checker (DOC)
10:58 < pinchartl> - Upstream DISCOM support
10:58 < pinchartl> - DAB/FM
10:58 < pinchartl> - Video codecs
10:58 < pinchartl> * Bug fixes
10:58 < pinchartl> - Dynamic BRS/BRU allocation in display pipelines
10:58 < pinchartl> - drmModeAtomicCommit() API test makes VGA side bad
10:59 < pinchartl> * Virtualization
10:59 < pinchartl> - Display virtualization (DU0 + DU1 on host, DU2 + DU3 on guest)
10:59 < pinchartl> that's it
11:00 < pinchartl> JTAG support for Gen3 is a core task
11:00 < pinchartl> any other proposal ?
11:00 < jmondi> pinchartl: I have received a D3 I assume for VIN parallel input support
11:00 < uli___> what exactly does "jtag support" mean?
11:01 < pinchartl> kbingham: that question is for you
11:01 < kbingham> uli___: The ability to use OpenOCD to connect and debug the kernel using GDB on Gen3 platforms
11:01 < pinchartl> jmondi: indeed, let me add that
11:01 < jmondi> and I feel like we should move gmsl in the open asap (as soon as rcar-vin and rcar-csi2 are mainlined)
11:01 < jmondi> we are actually piling too many pateches around
11:02 < pinchartl> there's also near-lossless compression upstreaming, I'll add that too
11:02 < jmondi> also, ov10635 driver should be separated from rdacm20, if we want to ease rdacm21 support...
11:03 < pinchartl> I'm not sure how to handle the rdacm21 yet
11:03 < pinchartl> but I agree about moving development to the open
11:03 < morimoto`> sorry, I need to go too.
11:03 < morimoto`> bye
11:03 -!- morimoto` [~user@relprex1.renesas.com] has left #periperi ["ERC Version 5.3 (IRC client for Emacs)"]
11:03 < pinchartl> morimoto`: no worries. have a nice evening
11:03 < pinchartl> dammsan: are you still there ?
11:04 < pinchartl> uli___: I think JTAG support means OpenOCD + gdb debugging for Gen3
11:04 < pinchartl> jmondi: we need to have upstreamable code then :-)
11:04 < jmondi> pinchartl: and code that works, also
11:05 < pinchartl> that an upstreaming requirement for me :-)
11:05 < jmondi> I feel like after Feb. code camp that should be our first effore for gmsl.. everything is so fragile (see moving to v4.15-rc1, we still have not been able to achieve)
11:05 < kbingham> uli___: I would particularly like to be able to use it - I have already hooked up a a BusBlaster - and it can see the 'dap' on the coresight module - but can't get access to the CPU's yet.
11:05 < jmondi> s/effore/effort/
11:05 < pinchartl> checking an older additional tasks list, there' also
11:05 < pinchartl> * DU interlaced modes support
11:05 < pinchartl> * VIN scaler
11:05 < pinchartl> * VIN TB/BT field support in continuous mode
11:05 < pinchartl> neg: or is that one complete already ?
11:05 < kbingham> uli___: If this is an area your familiar with - I would certainly be happy to hand the topic over to someone who knows more than me - I just want it to work :)
11:06 < uli___> i did that for kzm9g before...
11:06 < pinchartl> kbingham: what's the status of * ADV7482 CEC support ?
11:06 < neg> pinchartl: no VIN UDS scaler work not started, feels like the driver shuld be upstreamd first
11:06 < kbingham> pinchartl: Not yet done, and therefore also on that list :)
11:06 < pinchartl> neg: I meant TB/BT in continuous mode
11:07 < neg> pinchartl: VIN TB/BT in continuous mode would depend on the rework of the capture loop to not use single capture mode, which I hope to also be able to postpone to after Gen3 support is megred upstream
11:07 < pinchartl> neg: OK
11:08 < pinchartl> uli___: how about on your side, do e need more additional task for the GPU ? or is there something else you'd like to work on ?
11:09 < uli___> the jtag thing sounds good, actually
11:09 < pinchartl> I like it too, but it's a core task, so it has to be coordinated with Geert
11:09 < pinchartl> anything on the multimedia side ? :-)
11:09 < uli___> about the gpu, i'd like to finish the current task before i make up my mind about it
11:10 < pinchartl> when do you expect to finish it ?
11:10 < uli___> it's due on friday
11:11 < uli___> apart from that, dab/fm stuff sounds interesting, depending on the details
11:11 < kbingham> DAB/FM was just something I put on the list because I noticed there is some hardware somewhere in the R-Car to handle it.
11:11 < pinchartl> I don't know if it's a priority for Renesas, but we can certainly propose it
11:11 < kbingham> So I don't really know details beyond that.
11:14 < pinchartl> ok
11:15 < pinchartl> so I have a list of tasks that I will include in the report
11:15 < pinchartl> and we'll need to prioritize it, based on Renesas' feedback
11:16 < pinchartl> as neither Morimoto-san near Magnus seem to be here, I think we'll handle that through e-mail
11:16 < pinchartl> dammsan: still not available ?
11:16 < pinchartl> geertu: ping
11:18 < geertu> pinchartl: pong
11:19 < pinchartl> geertu: have we decided on whether the next meeting will be on the 9th or the 11th ?
11:20 < geertu> pinchartl: Not yet. wsa is unavailable on the 11th
11:20 < pinchartl> when do we decide ? :-)
11:21 < geertu> Does the 9th fit for all of you?
11:21 < pinchartl> works for me
11:22 < jmondi> that's fine!
11:22 < uli___> ok for me
11:22 < pinchartl> kbingham: neg: how about you ?
11:23 < kbingham> 9th works for me - probably better than the 11th for the same reasons as wsa  :D
11:24 < neg> works for me
11:24 < pinchartl> kbingham: what's happening on the 11th ?
11:25 < pinchartl> OK, let's go for the 9th then
11:25 < kbingham> pinchartl: It's the day after the day :)
11:25 < kbingham> I think I recall both me and WSA share the same birthday
11:26 < geertu> kbingham: You expect a hangover on the 11th?
11:26 < pinchartl> but isn't the 9th the day before the day then ?
11:26 < kbingham> geertu: Nope - I try hard to avoid hangovers lately ...
11:26 < geertu> kbingham: Screaming Hugo + hangover = profit?
11:27 < kbingham> geertu: haha
11:28 < pinchartl> so I think this is all for the multimedia meeting today
11:28 < pinchartl> does anyone else have anything to add ?
11:28 < neg> I'm good, thanks all
11:29 < pinchartl> thank you all for attending