path: root/wiki/Chat_log/20161123-mm-chatlog

Multimedia-chat-meeting-2016-11-23

09:03 < pinchartl> Magnus and Morimoto-san are excused
09:03 < pinchartl> it's a public holiday in Japan
09:03 < pinchartl> so let's get started
09:04 < pinchartl> first of all
09:04 < pinchartl> neither of you have reported anything by e-mail
09:04 < horms> Workers day iirc
09:04 < pinchartl> I could somehow see that coming as the time between the e-mails and the start of the meeting got shorter every time
09:05 < pinchartl> please fix that for next
09:05 < pinchartl> time
09:06 < kbingham> pinchartl: Just went out :D
09:06 < pinchartl> now that we're done with the scolding of the scoundrels, let's get really started :-)
09:07 < pinchartl> the only topic I have for today is status check for tasks
09:07 < pinchartl> does anyone want to add something to the agenda ?
09:08 < kbingham> Nope - perhaps today will be a quick meeting :D
09:08 < pinchartl> let's see
09:08 < pinchartl> you can start :-)
09:08  * kbingham walked into that one
09:08 < pinchartl> :-D
09:10 < kbingham> OK -A)  so I'm happy to say the BRU multiple input race has a fix - which makes me happy. tested over 1500 iterations, when normally it will occur in under 150. - I've also had a play with uli's HDMI patches ... just so I can actually see something out of the HDMI port at last. ... I found and fixed a small bug in raw2rgbpnm, and I've created tests for suspend resume using automated pm test framework.
09:11 < kbingham> And the best thing I've done this fortnight is my student loan is now clear \o/
09:11 < horms> hooray!
09:11 < pinchartl> congratulations :-)
09:11 < neg> :-)
09:11 < pinchartl> and good job
09:11 < kbingham> B) Next I will try to work through the VSP suspend resume issues, and do some final work on race's to see if I can find any more.
09:12 < kbingham> C) Nothing much holding me back - except for the hardware is hung and I have to figure out why. Probably getting itself misconfigured as a race waking up or clocks not being enabled in the right order ...
09:12  * kbingham wishes for a register that tells you why the hardware is hung :D
09:13 < geertu> kbingham: JTAG?
09:13 < kbingham> geertu: Aha - Yes - I did start to look at that too - but I haven't progressed as far as hooking up the blaster to the board :D (I was lacking in jumper cable connectors)
09:13 < pinchartl> geertu: it's not the system being hung, just the VSP
09:13 < pinchartl> kbingham: right ?
09:14 < pinchartl> or is it the full system ?
09:14 < kbingham> pinchartl: geertu: Yes - it's the hardware that is hung ... not the CPU.
09:14 < pinchartl> so just the VSP ?
09:14 < kbingham> - /hardware/VSP hardware/ : ;D yes
09:14 < pinchartl> ok
09:14 < pinchartl> I don't think JTAG would be very helpful then
09:15 < neg> kbingham: if you get JTAG running on Gen3, I'm interessted to hear about it :-)
09:15 < pinchartl> to get the HDMI output working, did you have to do anything else than applying the latest patch series ?
09:15 < kbingham> pinchartl: I agree - not in this instance ... Sorry - I thought geertu was asking as I had been talking with him about JTAG :D
09:16 < kbingham> pinchartl: No - I believe HDMI worked once I managed to get the patches applied correctly.
09:16 < pinchartl> ok, thank you
09:16 < pinchartl> next in alphabetical order, more
09:17 < pinchartl> since last meeting, I've followed up on the LVDS mode selection patches and posted a new version of the patch series
09:17 < pinchartl> or the two patch series I should say, covering both LVDS panel and LVDS encoder
09:18 < pinchartl> the feedback I got from Rob on the DT bindings was much more positive
09:18 < pinchartl> but Thierry Redding now chimed in and wasn't thrilled, to say the least
09:18 < pinchartl> we have very different opinions
09:19 < pinchartl> more work is needed to fix some issues
09:19 < pinchartl> and to either convince Thierry, or exhaust him until he stops caring
09:20 < pinchartl> a v3 is under preparation
09:20 < pinchartl> the LVDS encoder support series makes the DU driver cleaner by removing manual handling of bridges
09:21 < pinchartl> so I plan to make it a base for HDMI output upstreaming on Gen3
09:22 < pinchartl> I've also posted a few misc DU patches forward-ported from the BSP
09:22 < pinchartl> and they have been merged by Dave
09:22 < pinchartl> along with a bunch of other pending DU patches
09:23 < pinchartl> I've finished classifying the BSP patches
09:24 < pinchartl> I need to follow-up with the BSP team as it wasn't clear for some of them if they fixed real issues
09:24 < pinchartl> a few tasks have also been identified
09:24 < pinchartl> they will be candidate for additional tasks
09:25 < pinchartl> for the next two weeks
09:25 < horms> feel free to let me know if you want me to help contact the BSP team
09:25 < pinchartl> I still plan to post a proposal for the rotation and histogram API (haven't had time to address it yet)
09:25 < pinchartl> horms: thanks
09:26 < pinchartl> my plan is to ask Matsuoka-san, CC'ing the periperi mailing list
09:27 < horms> ok, that sounds like a good plan
09:27 < pinchartl> I will also start discussing V4L2 cache handling with upstream developer
09:28 < pinchartl> and upstream basic support for HDMI output on Gen3
09:29 < pinchartl> and, as I've already mentioned, still work on LVS panel and encoder support
09:29 < pinchartl> my main issues right now are Mauro doing his bad maintainer play again
09:30 < pinchartl> and Thierry not liking the LVDS panel DT bindings (and driver)
09:30 < pinchartl> that's it for me
09:30  * kbingham gets the popcorn to watch #v4l
09:31 < pinchartl> kbingham: it was quite surreal yesterday
09:31 < pinchartl> you need to know quite a bit about the technical problems being discussed to follow
09:31 < pinchartl> but I really felt I was in a parallel universe where the known rules of software architecture didn't apply
09:32 < kbingham> :)
09:32 < pinchartl> by the way the three new tasks I've identified from the BSP are
09:32 < pinchartl> - DU suspend/resume
09:33 < pinchartl> - Gen3 DU dot clock rate optimization (with DPLL)
09:33 < pinchartl> - Display interlace support
09:33 < pinchartl> the second one is part of Ulrich's HDMI output patches
09:33 < pinchartl> I might get it done as part of the HDMI output upstreaming, but might also leave it out
09:33 < pinchartl> depending on how much time it takes
09:33 < horms> Thanks, I'll try to add that to the google sheet if you haven't done so already
09:34 < pinchartl> do we have tasks in that spreadsheet ?
09:35 < horms> not as such.
09:35 < pinchartl> ok :-)
09:35 < horms> but we have "Features"
09:36 < pinchartl> I've added them to the multimedia todo list (will push after this meeting)
09:36 < horms> And I think it makes sense to map the Features to tasks where possible
09:36 < pinchartl> I'm done
09:36 < pinchartl> neg: you're next
09:36 < neg> a) Nothing
09:36 < neg> b) Address CSI2 review comments
09:36 < neg> c) None
09:36 < neg> EOT
09:38 < pinchartl> I assume you're busy with I/O and/or core tasks at the moment ?
09:39 < neg> Not super busy, trying to start a new core task and finish up my IO task so if yout got something that needs to be done I might have time
09:39 < pinchartl> I'm looking at the todo list
09:40 < pinchartl> we have a bunch of tasks planned for v4.10 that need to be pushed back
09:40 < pinchartl> for you that's
09:40 < pinchartl> ADV7482,v4.10,plan,niklas,Gen3 support upstream
09:40 < pinchartl> ADV7482,v4.10,plan,niklas,Interlace support upstream
09:40 < pinchartl> VIN,v4.10,public,niklas,Gen3 support
09:40 < pinchartl> VIN,v4.10,plan,niklas,Scaler support (on Gen3)
09:40 < pinchartl> VIN,v4.10,public,niklas,CSI2 interlace support upstream (Gen3)
09:40 < pinchartl> VIN,v4.10,public,niklas,CSI2 support upstream (Gen3)
09:40 < pinchartl> VIN,v4.10,plan,niklas,Gen3 support upstream (without CSI-2)
09:40 < pinchartl> should I push them back to v4.11 or directly to v4.12 ?
09:41 < neg> I don't think we can move forward with CSI2 without VIN and so far there have not been much review on the series which is huge so I assume it will miss v4.11, or how do you feel about it?
09:42 < pinchartl> I agree with you
09:42 < pinchartl> unless we decide it's a high priority and put lots of effort on that
09:42 < pinchartl> but that doesn't seem to be the case at the moment
09:42 < pinchartl> so v4.12 seems more realistic to me
09:43 < neg> for ADV7482 I have mixed feelings about having it in periperillist, is that something we still feel is a task to bring to upstrem since there are a huge effort to do so?
09:43 < pinchartl> I don't think so at the moment. I'll mark it as noplan
09:43 < pinchartl> but keep it there as it would be nice to have
09:45 < neg> yes I would love to get it upstream one day, I will keep to work on the driver so I can test VIN but the big push to make it ready for upstream I feel is a low prio thing
09:45 < pinchartl> thanks
09:45 < pinchartl> next, uli___ 
09:45 < uli___> as mentioned above, i have pushed out a base for upstreaming hdmi out on gen3
09:46 < uli___> and a tool for testing analog audio remotely with a loopback cable
09:46 < uli___> https://github.com/uli/atest
09:46 < uli___> for interested parties
09:46 < uli___> mixer settings are tricky, but the readme contains a few hints
09:46 < kbingham> uli___: I like atest - that's neat :D
09:46 < uli___> thank you :)
09:47 < uli___> next, i will upport/test/fix ipmmu patches for du, vsp and fcp
09:47 < uli___> haven't looked into that yet
09:47 < uli___> no issues
09:47 < uli___> that's it from me
09:48 < pinchartl> what's the status of VIN,v4.10,public,ulrich,Gen2 VIN integration ? have the patches been merged ?
09:48 < uli___> i'm not sure, actually
09:48 < pinchartl> horms might know ?
09:49 < uli___> or neg
09:49 < pinchartl> those are DT patches
09:49 < horms> I don't believe so
09:49 < horms> I can look up but most likely I was expecting some Acks and then it fell off my radar
09:50 < pinchartl> I see the following patches in your tree
09:50 < pinchartl> 06b64afa6e98 ARM: dts: r8a7793: Enable VIN0-VIN2
09:50 < pinchartl> 84e3a74664c5 ARM: dts: koelsch: add HDMI input
09:50 < pinchartl> 56548d0c5aea ARM: dts: lager: Add entries for VIN HDMI input support
09:50 < pinchartl> uli___: were there more ?
09:50 < uli___> for gose, i think there was an issue
09:50 < uli___> it's missing the enablement for analog and hdmi in the dts file
09:51 < uli___> i'll look into that
09:51 < pinchartl> thanks
09:51 < horms> yes, i think those things are missing from gose
09:51 < horms> i'm entirely usnure why
09:51 < pinchartl> I'll update the target to v4.11 then
09:52 < horms> please feel free to ping me any time if dt patches seem stuck
09:53 < pinchartl> "[PATCH v2 2/3] ARM: dts: gose: add HDMI input" is ready to be merged with a small fix that I requested
09:53 < pinchartl> "[PATCH v2 3/3] ARM: dts: gose: add composite video input" will require a bit more work
09:54 < uli___> yes, i think i'll resend the hdmi part individually, so it won't me tainted by disparaging comments :)
09:54 < uli___> s/me/be/
09:55 < pinchartl> you can also start to update the ADV7180 DT bindings already
09:55 < pinchartl> the only open item is what compatible strings to use
09:55 < pinchartl> the rest can already be fixed
09:56 < uli___> i'll have a look
09:57 < pinchartl> thanks
09:58 < pinchartl> next meeting
09:58 < pinchartl> I propose 2016-12-07 at the same time as today
09:58 < pinchartl> 08:00 GMT / 09:00 CET / 10:00 EET / 17:00 JST
09:58 < neg> uli___: thanks for all your work for VIN, I'm looking forward to v4.10 and out of the box HDMI VIN for gen2 :-)
09:59 < neg> Meeting time works for me
09:59 < uli___> you are welcome :)
09:59 < kbingham> uli___: Then you can write a vtest to do a video loopback with embedded data verification too :D
10:00 < horms> FWIIW that is not a holiday in Japan
10:00 < uli___> kbingham: don't give anybody ideas
10:00 < pinchartl> horms: I'm sure they'll find another excuse ;-)
10:00 < pinchartl> (it's the day after the Finnish independence day)
10:00 < horms> no comment
10:00 < neg> vtest +1 :D
10:00 < pinchartl> does anyone have anything to add ?
10:00 < kbingham> uli___: Should be possible to do straight byte comparisons - but encoding the bytes sounds more fun :D
10:00 < horms> Indepemdence from Sweeden_
10:00 < horms> _
10:00 < horms> ?
10:01 < pinchartl> from Russia I believe
10:01 < horms> Ok. I think they may bave become independent a few times. But I could easily be wrong.
10:02 < pinchartl> :-)
10:02 < pinchartl> if there's no more comment I propose adjourning this meeting. does anyone second ?
10:02 < neg> Any multimedia meeting plan for FOSDEM timing I should keep in mind when bookig my trip?
10:02 < pinchartl> neg: good point
10:02 < kbingham> Ah yes- I'll be at FOSDEM too - would probably like to book flights early
10:02 < pinchartl> it looks like there will be a core and I/O day on Friday before the FOSDEM
10:03 < pinchartl> it's not clear yet whether we'll need a multimedia meeting, but it will likely be useful to discuss additional tasks at least
10:03 < pinchartl> so I propose reserving time for that on Thursday
10:04 < pinchartl> I'll be in Belgium anyway :-)
10:04 < pinchartl> so it's easy for me
10:05 < pinchartl> but if any of you don't want to spend an extra day in Brussels needlessly, we need to decide whether to organize a meeting or not
10:05 < pinchartl> I assume everybody would arrive on Thursday at the latest anyway, to be there for the core and I/O meetings on Friday
10:06 < pinchartl> so if you book a flight that doesn't arrive too late and we host the meeting in the afternoon then we should be good
10:06 < pinchartl> any thoughts on that plan ?
10:06 < neg> For me Thursday is fine, maybe have a afternoon meeting so ppl can arive that day and still not be late for meeting?
10:06 < pinchartl> yes, that's the idea
10:07 < kbingham> Looks like my light will get in at 12:50 local time ...
10:07 < kbingham> how long does it take to get fro m the airport to the centre.
10:07 < pinchartl> the train takes a bit less than 20 minutes
10:08 < pinchartl> but you also need to get out of the plane and to the train station :-)
10:08 < pinchartl> you might be a bit late for lunch, but for an afternoon meeting it's definitely fine
10:08 < pinchartl> uli___: how about you, what are you travel plans ?
10:08 < uli___> i can't confirm that i will be at fosdem at all yet...
10:08 < uli___> so don't consider me when planning any meetings
10:09 < uli___> judging from the train schedule
10:09 < uli___> i'd be there at 5 something
10:10 < pinchartl> 5pm I assume
10:10 < uli___> yes
10:11 < pinchartl> ok, so we should decide as soon as possible whether to organize a meeting on Thursday then
10:12 < pinchartl> I'll ask Magnus and Morimoto-san about their opinion in the meeting report and we'll then decide
10:12 < kbingham> at the very least we could do dinner on thursday evening
10:12 < horms> that sounds like a fine idea to me :)
10:12 < pinchartl> yes dinner is certainly a must :-)
10:13 < pinchartl> if there's no more comment I propose adjourning this meeting. does anyone second ?
10:14 < neg> second
10:14 < pinchartl> thank you
10:14 < pinchartl> the meeting is adjourned, thank you everybody for attending