h1. chat report 2016-06-22 Attendees: - Kieran - Laurent - Magnus - Morimoto-san - Niklas - Ulrich Topic 1. Status check for the multimedia tasks ---------------------------------------------- - FDP,v4.8,prototype,kieran,Develop and upstream driver A first version has been posted to the linux-media mailing list. Since then, the driver is now adapted to using a job queueing system to allow abstracting fields from v4l2 buffers, and de-interlacing tested with PREVFIELD, NEXTFIELD, FIXED2D. 3D deinterlacing is in progress, it requires tracking video buffers between m2m job runs. Gstreamer does not (yet) support asynchronous input/output which means testing is not as elegant as for the earlier stages. V4l2-compliance tests pass without failure. A new version of the patches is expected for this weekend. - DU,?,plan,laurent,IPMMU integration on Gen3 - DU,?,plan,laurent,IPMMU support on Gen3 (through VSPD+FCP) No progress. - VSP,v4.7,merged,laurent,Plane alpha + pixel alpha (on Gen3) Patches have been merged in the linux-media tree. - VSP,?,plan,laurent,Fixed alpha support (VI6_DPR_*_ROUTE.FXA) - VSP,?,plan,laurent,UDS regression fix - VSP,v4.8,plan,laurent,Fix suspend/resume crash No progress. The latter will not make it to v4.8. Setting the target back to ? as the job hasn't been scheduled yet. - VSP,v4.8,public,laurent,CLU 2D and 3D mode support - VSP,v4.8,public,laurent,CLU WARN_ON fix - VSP,v4.8,public,laurent,CLU/LUT upstream API - VSP,v4.8,public,laurent,CLU/LUT support submitted upstream on Gen3 - VSP,v4.8,public,laurent,HGO operation mode selection - VSP,v4.8,public,laurent,HGO support upstream on Gen3 - VSP,v4.8,public,laurent,HGO test application Patches are public, waiting a few more days for review before sending a pull request for upstream merge. A pull request has been sent for renesas-drivers already. - VSP,v4.8,public,laurent,HGO test application HGO tests are included in http://git.ideasonboard.com/renesas/vsp-tests.git. They currently require to use a special branch of yavta which will be merge into master once the HGO kernel patches are accepted. The same apply to the LUT/CLU tests. Every new feature added to the VSP driver now comes with related test cases. - VSP,v4.8,public,laurent,Horizontal/vertical flip support New task. The patches are public, waiting a few more days for review before sending a pull request for upstream merge. A pull request has been sent for renesas-drivers already. - VSP,v4.8,prototype,laurent,Rotation support New task. A prototype exists, will be cleaned up and posted, hopefully today. Rotation doesn't work properly, it corrupts the output image. I will send a report describing the issue in detail and Morimoto-san will share it with the BSP and/or HW team. - VSP,?,prototype,laurent,V4L2 request API upstream - VSP,v4.8,public,laurent,V4L2 request API usable prototype An updated patch series is currently being tested with the new MC test tool developed by Intel. Test instructions are being written. The patches will be posted along with test scripts and test instructions, the target is Monday. Both Intel and Google have expressed interest for the request API, I expect them to help pushing for the request API to be accepted upstream. - DU,v4.9,plan,laurent,Fix 3 planes formats New request from the BSP team. Support for formats with 3 planes is broken. Setting the target to v4.9, but the fix might come earlier. - AUDIO-DMAC,?,plan,magnus,IPMMU integration on Gen2 - AUDIO-DMAC,?,plan,magnus,IPMMU integration on Gen3 - AUDIO-DMAC,?,plan,magnus,IPMMU support on Gen2 - AUDIO-DMAC,?,plan,magnus,IPMMU support on Gen3 No progress - VIN,?,plan,magnus,IPMMU integration on Gen2 - VIN,?,plan,magnus,IPMMU integration on Gen3 - VIN,?,plan,magnus,IPMMU support on Gen2 - VIN,?,plan,magnus,IPMMU support on Gen3 A local prototype exists, setting the status to prototype. It should be posted in the not too distant future. - RSND,2016-06-30,public,morimoto,DT bindings for graph sound - RSND,2016-06-30,prototype,morimoto,dw-hdmi-ahb-audio prototype on Gen3 - RSND,2016-06-30,prototype,morimoto,HDMI SSI prototype on Gen3 - RSND,2016-06-30,plan,morimoto,HDMI sound Upstream support without hotplug on Gen2 The clean up patches needed to develop OF graph sound support are still under review. The maintainer has been pinged but hasn't replied yet. Work is now focused on the next step of OF graph support. Simple OF graph for ALSA is fine, but the complex features are more challenging. Updating the target to v4.10. - RSND,2016-09-30,plan,morimoto,Hotplug support upstream on Gen3 No progress. - ADV7482,v4.7,plan,niklas,Prototype on Gen3 It turns out work was needed in the V4L2 framework to support the ADV7482. Patches have been posted for that, and the rcar-vin driver is being updated accordingly. The ADV7482 prototype should be posted on Monday and the updated rcar-vin driver soon after that. The prototype won't support controls in the first version as that will require more work in the V4L2 framework. - VIN,v4.7,plan,niklas,CSI2 prototype (Gen3) The CSI-2 prototype is done and should be posted on Monday (pending possible modifications needed by the rework of the ADV7482). - VIN,v4.8,plan,niklas,Gen3 support The code has been posted publicly. - VIN,v4.8,plan,niklas,Scaler support (on Gen3) - ADV7482,v4.9,plan,niklas,Gen3 support upstream - ADV7482,v4.9,plan,niklas,Interlace support upstream - VIN,v4.9,plan,niklas,CSI2 interlace support upstream (Gen3) - VIN,v4.9,plan,niklas,CSI2 support upstream (Gen3) - VIN,v4.9,plan,niklas,Gen3 support upstream (without CSI-2) No progress. - DU,v4.7,plan,ulrich,Atomic API test program The task is complete. The elinux.org wiki will be updated today. - DU,v4.7,public,ulrich,HDMI output on Gen3 prototype The task is complete. An updated version with a fix for ES1.1 will be posted. - DU,v4.7,prototype,ulrich,Test setup with HDMI output to HDMI input loopback (without EDID) The task is complete. - DU,v4.7,public,ulrich,EDID generation support for the HDMI loopback test setup No progress, documentation will be written before end of June. - VIN,v4.8,public,ulrich,Add DV timings support to rcar-vin - DU,v4.9,plan,ulrich,HDMI output on Gen3 upstream No progress. h1. chat log morning [16:54] hi good morning or evening depending on where you are :-) Yes, here is evening :) [16:55] sorry I'm a bit late [16:57] hello everybody [16:58] will Magnus join us today ? hi, Magnus is here perfect, let's start (taking my notes) topics for today [16:59] Topic 1. Status check for the multimedia tasks Topic 2. Additional '50%' tasks (June) Topic 3. Additional '50%' tasks (Q3) Topic 4. Next meeting [17:00] anything else ? nothing from me :) ok [17:01] let's go for the alphabetical order again [17:02] Kieran told me he might not be able to join today kbingham: are you there ? FDP,v4.8,prototype,kieran,Develop and upstream drive [17:03] r a first version of the driver is public more work has gone into supporting the different deinterlacing modes including some internal refactoring related to that almost here - but on a very dodgy connection :( [17:04] no worries when do you expect to post a v2 ? FDP1 - driver now adapted to using a job queueing system to allow abstracting fields from v4l2 buffers, and de-interlacing tested with PREVFIELD, NEXTFIELD, FIXED2D. Need to track VB痴 between m2m job runs to finish the 3d deinterlacing. Gstreamer does not (yet) support asynchronous input/output which means testing is not as elegant as for the earlier stages. V4l2-compliance tests pass without failure. [17:05] but yes - what laurent said :) when do you expect to post a v2 ? [17:07] I'm hoping at the weekend (although now that is going to be the end of the weekend :) ) ok :-) [17:08] my turn then DU,?,plan,laurent,IPMMU integration on Gen3 DU,?,plan,laurent,IPMMU support on Gen3 (through VSPD+FCP) no progress VSP,?,plan,laurent,Fixed alpha support (VI6_DPR_*_ROUTE.FXA) [17:09] VSP,?,plan,laurent,UDS regression fix VSP,v4.8,plan,laurent,Fix suspend/resume crash no progress the latter will not be ready for v4.8 I'll set the target version back to ? as the job hasn't been scheduled yet [17:10] VSP,v4.8,public,laurent,CLU 2D and 3D mode support VSP,v4.8,public,laurent,CLU WARN_ON fix VSP,v4.8,public,laurent,CLU/LUT upstream API VSP,v4.8,public,laurent,CLU/LUT support submitted upstream on Gen3 VSP,v4.8,public,laurent,HGO operation mode selection [17:11] VSP,v4.8,public,laurent,HGO support upstream on Gen3 patches are public, I'll wait a few more days for review and send a pull request [17:12] 1 request from BSP they want to have CLU 2D/3D mode on renesas-driver I've sent a pull request to Geert for renesas-drivers OK, nice [17:13] VSP,v4.8,public,laurent,HGO test application HGO tests are included in http://git.ideasonboard.com/renesas/vsp-tests.git [17:14] they currently require to use a special branch of yavta I will merge that branch back to master once the HGO kernel patches are accepted same for the LUT/CLUT tests [17:15] every new feature I add to the VSP driver now includes corresponding test cases VSP,v4.8,public,laurent,Horizontal/vertical flip support [17:16] that's a new task the patches have been posted as well and are included in the renesas-drivers pull request [17:17] VSP,v4.8,public,laurent,Rotation support new task as well [17:18] I haven't published the patch yet (the status should be prototype, not public) I'll try to do so today but it doesn't work properly :-/ http://www.ideasonboard.org/vsp-rotation.png [17:19] that's what I get after rotation http://www.ideasonboard.org/vsp-rotation-expected.png [17:20] is the expected result I'm stuck at this point and suspect a hardware problem morimoto: I'll send you a mail with a description of the problem after posting the patch could you forward it to the BSP team ? [17:21] OK, I will ask it to BSP/HW team I'd like to know if they have tested rotation on H3 thank you VSP,?,prototype,laurent,V4L2 request API upstream [17:23] VSP,v4.8,public,laurent,V4L2 request API usable prototype after rotation that's the only task I have left for this month I have an updated patch series I'm testing it with the new MC test tool developed by Intel [17:24] I'm also writing down the test instructions and will post the patch series with the instructions and test scripts my target is Monday that's all for me [17:25] morimoto: any other request from the BSP team ? they want to know about request API status [17:26] Ahh, and we got pixsel format request from BSP team I think I've just addressed that question. another note on this topic, both Intel and Google are interested in the request API, so I expect to be pushed by them too I posted it to this Multimedia meeting invitation [17:27] You addressed which one ? request API ? pixsel format ? request API. the target for the new patch series with the test scripts and test instructions is Monday [17:28] let me check my e-mails [17:29] OK, thanks when did you post it ? [17:30] right now ? ah no on Friday Yes, Friday it includes 2 request [17:31] not a big deal, but we can get BSP team smile :) I'll fix that [17:32] gr8 And we got request from BSP team, but it is very big topic when do they need it ? [17:33] it can be next PeriPeriCon (!= MiniPeriCon) topic (the 3 planes formats) They already have patch on locally, so whenever is OK ok I'll see if I can fix that for v4.8 but it might be too short, there's only a few weeks left [17:34] ok, no problem what's the new very big topic ? They want to use VSP without cache [17:35] -> no copy where is there a copy ? ERC> /no copy/zero copy/ *** no: Unknown command and where does the VSP use case ? s/no copy/zero copy/ buffers are mapped with uncached mappings and video frames are not copied I forget detail, but they have something. [17:36] Anyway, it can be PeriPeriCon topic can you find out what this is about and send an e-mail with more information ? OK, will do [17:37] thank you next, dammsan AUDIO-DMAC,?,plan,magnus,IPMMU integration on Gen2 AUDIO-DMAC,?,plan,magnus,IPMMU integration on Gen3 AUDIO-DMAC,?,plan,magnus,IPMMU support on Gen2 AUDIO-DMAC,?,plan,magnus,IPMMU support on Gen3 VIN,?,plan,magnus,IPMMU integration on Gen2 VIN,?,plan,magnus,IPMMU integration on Gen3 VIN,?,plan,magnus,IPMMU support on Gen2 VIN,?,plan,magnus,IPMMU support on Gen3 no progress ? [17:38] Please wait. MagPeri is fighting his headache hey guys [17:39] hi Magnus FYI, i've prototyped VIN bits with local code now on Gen2 and Gen3 so they can maybe move from plan to local if there is such thing? [17:40] no progress about AUDIO-DMAC nice, thanks there's no such thing as local, maybe prototype ? [17:41] yeah, sure the code is not public yet [17:42] but i intend to send it out in the not so distant future i saw that the IOMMU maintainer pinged about how to merge so i need to act on that anything else? if not i'll head back into paper work land =) that's all, good luck with the paperwork [17:43] oh, I forget about this task VSP,v4.7,merged,laurent,Plane alpha + pixel alpha (on Gen3) patches have been merged (in the linux-media tree) [17:44] next, morimoto RSND,2016-06-30,public,morimoto,DT bindings for graph sound [17:45] RSND,2016-06-30,prototype,morimoto,dw-hdmi-ahb-audio prototype on Gen3 RSND,2016-06-30,prototype,morimoto,HDMI SSI prototype on Gen3 RSND,2016-06-30,plan,morimoto,HDMI sound Upstream support without hotplug on Gen2 RSND,2016-09-30,plan,morimoto,Hotplug support upstream on Gen3 I posted clean up patch which is 1st step for OF graph sound support it is still under review I sent ping mail to maintainer, but no response yet [17:46] Now, I'm development next step ALSA sound have many features. I'm working for this "many features" with OF graph simple OF graph + sound is already OK [17:47] I'm fighting with "complex features" + OF graph now ALSA need to more of_graph_xxx() function/macro :) over from me thank you [17:48] the first 4 tasks have 2016-06-30 as their target date should we update that ? I think so If you don't give me penalty :) [17:49] I'll let you sort that out with the BSP team :-) what should be the new target ? maybe a kernel version instead of a date, to match the other tasks ? [17:50] It seems this clean up patch-set itself (= 1st step) taks long term... s/taks/takes/ and OF graph patch will taks more time... [17:51] +3 month ? that's fine with me Thanks v4.10 then ? [17:52] Could you add above "pixel" tasks to peripelist ? pinchartl: Yes nice I've added "DU,v4.9,plan,laurent,Fix 3 planes formats" [17:53] OK, nice BTW, you can use "make ver" on peripelist if you want to have version base output thanks [17:54] neg: your turn ADV7482,v4.7,plan,niklas,Prototype on Gen3 ADV7482,v4.9,plan,niklas,Gen3 support upstream ADV7482,v4.9,plan,niklas,Interlace support upstream VIN,v4.7,plan,niklas,CSI2 prototype (Gen3) VIN,v4.8,plan,niklas,Gen3 support VIN,v4.8,plan,niklas,Scaler support (on Gen3) VIN,v4.9,plan,niklas,CSI2 interlace support upstream (Gen3) VIN,v4.9,plan,niklas,CSI2 support upstream (Gen3) VIN,v4.9,plan,niklas,Gen3 support upstream (without CSI-2) CSI2 is done I hope, but have not posted it yet if something comesup with the rework for adv7482 but will post it soon [17:55] adv7482 is started but it turnd out work was needed on the v4l framework, I have posted patches for that [17:56] right now I'm updating the rcar-vin driver to use the new v4l feature needed to use adv7482 and after that I will repost the rcar-vin driver also addressing your review comments [17:57] plan is to post csi-2 and adv7482 prottypes on monday and the updated rcar-vin soon after that [17:58] I plan to not support controlls for the adv7482 in the prototype since more work is needed in v4l to support that for our use case VIN,v4.8,plan,niklas,Gen3 support is public [17:59] do you think it will make it to v4.8 ? [18:00] (Gen3 support for VIN, not ADV7482 and CSI-2- I think I can get the patches updated in time but I think review is going to drag out due to the complexity needed to support the runtime input selection [18:01] so I do not feel confident about that [18:02] ok, v4.9 then [18:03] same for scaler support ? ah wait "VIN,v4.8,plan,niklas,Gen3 support" isn't for upstream so v4.8 is fine yes same for scaler, I really would like reviews on the group concept before I add scalers to the mix making it even more complex [18:04] ok [18:06] regarding controls for ADV7482 have you seen Hans' proposal ? only what he posted in #v4l that's what I was referring to [18:07] ok, thanks for the report next, uli___ DU,v4.7,plan,ulrich,Atomic API test program DU,v4.7,public,ulrich,HDMI output on Gen3 prototype DU,v4.7,prototype,ulrich,Test setup with HDMI output to HDMI input loopback (without EDID) DU,v4.7,public,ulrich,EDID generation support for the HDMI loopback test setup VIN,v4.8,public,ulrich,Add DV timings support to rcar-vin [18:08] DU,v4.9,plan,ulrich,HDMI output on Gen3 upstream no progress, basically nothing for public consumption, at any rate how about "DU,v4.7,plan,ulrich,Atomic API test program" ? [18:09] same as last time, i.e. it's out i didn't move the wiki pages, though :) sorry about that could you please do so today, and link to the page from http://elinux.org/R-Car/Devices ? [18:10] ok thank you regarding "DU,v4.7,public,ulrich,HDMI output on Gen3 prototype" [18:13] what's the status of the latest code ? and do you plan to send an update ? yes, a small one, with the fix for es1.1 that's nice for me thanks [18:14] I'll mark the task as complete We will have other solution for this kind of ESx support. it can be MiniPeriCon topic i won't be able to attend; please keep me up-to-date on that [18:15] Oops, indeed I created MiniPeriCon / RenesasCon wiki page [18:16] about "DU,v4.7,prototype,ulrich,Test setup with HDMI output to HDMI input loopback (without EDID)", is there documentation available in the elinux.org wiki ? you can check log pinchartl: none yet [18:17] do you plan to add that this quarter ? (meaning in the next 8 days :-)) [18:18] yes, i consider it part of the add. task thank you that's it for the tasks check [18:19] Topic 2. Additional '50%' tasks (June) we have - VIN CSI-2 Support Prototype - - ADV7482 Driver Prototype - VSP CLU Improvements - VSP Rotation Support - HDMI Input/Output Loopback for R-Car Gen2 [18:20] VIN CSI-2 on track will be posted latest on monday (scratch the last one, bad copy & paste) [18:21] ADV7482 have local prototype but want to do more work to only have one device for both the 1-lane and 4-lane outputs hope to be done and post it on Monday (no, don't scratch it, it was the right copy & paste :-)) [18:22] if I find that more work is needed in the v4l framework to support this I will post my local prototype - FDP Support on Gen3 [18:23] neg: thanks I see that both prototypes will be posted for the end of the month, that's enough for me on my side VSP CLU is ready and VSP rotation partially implemented as explained before [18:24] h/v flip works, rotation doesn't we've already discussed HDMI loopback and FDP support so we seem to be on track Topic 3. Additional '50%' tasks (Q3) [18:25] Q3 will be split in two batches we need to agree on the first batch before the end of this month and we'll discuss the second batch during peripericon [18:26] we have proposed tasks for the first batch I've already discussed them with neg uli___: what's your opinion about working on the following ? - VIN HDMI input EDID on Gen2 (1 week) - Gen2 VIN integration (1 week) [18:27] the first one is about replacing the hardcoded EDID that Hans complained about with something fit for upstream the second one is quite light and is about integrating VIN in DT for all the Gen2 boards for what period of time would that be? both are upstream-focussed [18:28] they are due for mid-August you can start whenever you want :-) i see i'll have to check with geert and wolfram first, but generally i'd be fine with that so you have 6 weeks allocated for additional tasks in Q3 [18:29] yes 3 weeks per batch those two tasks cover two weeks, there would be one week for core or I/O if Geert or Wolfram need you for more than a week we can possibly trim it down a bit on the multimedia side [18:30] I'd like to keep VIN HDMI input on Gen2 but dammsan might have a different opinion yes, i'd prefer that, too so, assuming there's no issue with time allocation between groups, the above seems ok to you ? [18:31] i'm ok with the tasks, yes thank you kbingham: are you still there ? pinchartl: yes, I'm here pinchartl: different opinion? =) [18:32] dammsan: on what you think is more important in case Ulrich needs to spend more time on core or I/O and less on multimedia kbingham: I've started discussing tasks and budget for Q3 for you with Magnus, this is still ongoing, I'll get back to you ASAP about that [18:33] pinchartl: i just think that simon is good at doing integration, and ulrich has good hacking skill but you'll figure it out =) i'm fine with current setup [18:34] dammsan: so we have the same opinion, perfect yes pinchartl:. ok Topic 4. Next meeting [18:35] keeping the current bi-weekly schedule, that would be the 6th of July same time as today would that work for everybody ? works for me [18:36] works for me. And I won't be on the end of a piece of string either :) me, too :-) [18:37] kbingham: let me know when your internet connection will allow you to discuss new tasks it works for me too dammsan: for you too ? [18:38] I'll assume that's a yes [18:39] that's all for today pinchartl: Ack. thank you everybody for attending don't forget the reports for this month I know you all love them it always goes with an invoice; that works for me :) [18:40] pinchartl: do you want copies of our reports so you too can enjoy them :-) neg: I'll pass this time, thank you :-) neg: I'll pass this time, thank you :-) [18:41] (oops) pinchartl: unfortunately I can't report to you :) I'll enjoy the elinux.org wiki only morimoto: you can if you really insist, and I promise I won't read the report in that case :-) (report and invitation for next meeting sent) [18:42] pinchartl: hehe OK [18:43] So, can I go back to home ? [18:53] you can go wherever you want :-) [18:54] Hehe, OK. Thank you for chat meeting. [18:55] Bye Bye