path: root/wiki/Chat_log/20170621-mm-chatlog

Multimedia-chat-meeting-2017-06-21

09:03 < pinchartl> topics for today are
09:03 < pinchartl> - status update
09:03 < pinchartl> - additional tasks for Q3 (I'll keep this very brief)
09:03 < pinchartl> - next meeting
09:03 < pinchartl> anything else ?
09:04 < pinchartl> ok
09:04 < pinchartl> Topic 1. Status check for the multimedia tasks
09:04 < pinchartl> jmondi: you can start
09:05 < dammsan> hi guys sorry about the delay
09:05 < jmondi> pinchartl: thanks
09:05 < pinchartl> hi Magnus !
09:05 < jmondi> I guess this is punishment for having sent a late update
09:05 < jmondi> A)
09:05 < jmondi> - OSS Japan]
09:05 < pinchartl> jmondi: not quite, it's alphabetical order. you can blame your first name ;-)
09:06 < jmondi> pinchartl: damn parents!
09:06 < jmondi> - digital/parallel input on Gen3
09:06 < jmondi> -- hw setup (which took me longer than expected for several reasons)
09:07 < jmondi> -- sent RFC patches to add support for digital input on Gen3
09:07 < jmondi> B) resume 8 channel camera setup where we left it in Japan
09:07 < jmondi> - Address review comments for digital input on Gen3
09:07 < jmondi> C = NULL
09:07 < jmondi> --eot--
09:08 < pinchartl> so the hardware setup for parallel camera is now working ?
09:08 < jmondi> meh
09:08 < jmondi> I can read/write registers (chip ID excluded)
09:09 < pinchartl> can you read other registers ?
09:09 < jmondi> on the I2c bus I mean
09:09 < jmondi> yes
09:09 < pinchartl> or do they all return 0 ?
09:09 < jmondi> no, just chip ID
09:09 < pinchartl> so other registers return meaningful values, but chip ID returns 0 ?
09:09 < jmondi> which is weird
09:09 < jmondi> yep
09:09 < pinchartl> that's very weird
09:09 < jmondi> it is
09:09 < pinchartl> there's something fishy there
09:10 < geertu> They forgot to program the chip with the chip ID? Can you write to the chip ID register (once)?
09:10 < pinchartl> jmondi: have you tested video capture ? does it work ?
09:10 < jmondi> to clear things out, I have planned to not only trust i2c-gpio debug ouput, but use a logica analyzer (which I now have) and see what's actually on the bus
09:11 < pinchartl> jmondi: good idea
09:11  * neg is lurking from cellphone
09:11 < pinchartl> hi neg 
09:11 < jmondi> pinchartl: nope, I have problems with media-controller complaining about broken pipe when sending IOCTLs on the video device node
09:11 < pinchartl> jmondi: ok
09:11 < jmondi> geertu: no, it's read only
09:13 < pinchartl> jmondi: thank you
09:13 < pinchartl> next, kbingham[m] 
09:14 < pinchartl> I'll copy & paste his report
09:14 < pinchartl> since last meeting
09:14 < pinchartl> - OSS Japan
09:14 < pinchartl> - ADV748x v4, v5
09:14 < pinchartl> The final major blocker for upstream integration was the lack of EDID  support. Hans would not accept upstream without this, so EDID has been implemented and tested it as working. (Can now capture from Laptop output). Once Hans' remaining comments are resolved a (final?) v6 to be posted very  soon.
09:14 < pinchartl> - Discovered a bug in "v4l: vsp1: Repair suspend resume operations for video pipelines"
09:14 < pinchartl> The bug was introduced from the HGO/HGT control lock. Full log is available at http://paste.ubuntu.com/24864203/
09:14 < pinchartl> - Supported Laurent in preparing VSP-DU branch for upstream pull
09:14 < pinchartl> - Tested Sakari's ACPI Graph / fwnode branches (used by ADV748x)
09:14 < pinchartl> for the next two weeks:
09:14 < pinchartl> - v6 ADV748x hopefully for mainline integration (DT bindings approval still pending)
09:14 < pinchartl> - Investigate HGO/HGT locking issue.
09:14 < pinchartl> - Work with Hans to get remaining VSP1 patches integrated upstream
09:14 < pinchartl> issues and blockers
09:14 < pinchartl> - HGO/HGT issue hampered desires to have the pending VSP1 patches upstream
09:14 < pinchartl> already, but should not block progress. It should be possible to split those patches out from testing and get some more momentum, and/or repair the issue.
09:15 < pinchartl> next, me
09:15 < pinchartl> since last meeting
09:15 < pinchartl> - OSS Japen
09:15 < pinchartl> Japan even
09:15 < pinchartl> - Holidays
09:15 < pinchartl> - H3 ES2.0 display support
09:16 < pinchartl> I've discovered issues in the patches I've posted and I'm working on addressing them
09:16 < pinchartl> - M3-W HDMI output support
09:16 < pinchartl> This is useful to test the ES2.0 code as the M3-W VSP and DU are similar to the H3 ES2.0
09:16 < pinchartl> for the next two weeks
09:17 < pinchartl> - Continue H3 ES2.0 and M3-W display-related work
09:17 < pinchartl> - VSP / DU initialization order fix
09:17 < pinchartl> issues and blockers
09:18 < pinchartl> - Salvator-XS stuck in customs and Salvator-X H3 ES2.0 unavailable through remote access, so I'm left without a way to test H3 ES2.0 code
09:19 < dammsan> pinchartl: i'm working on enabling remote access for you
09:19 < pinchartl> dammsan: thank you. I'll also need local access very soon
09:19 < pinchartl> the shipping company hasn't replied to my e-mails
09:19 < pinchartl> possibly because they don't speak English
09:20 < jmondi> that's a good start with them, indeed
09:21 < pinchartl> or it might be that all my e-mails end up in their spam box
09:22 < pinchartl> I've sent them e-mails on Monday and Tuesday and they haven't replied
09:23 < pinchartl> dammsan: H3 ES2.0 will very likely not be ready in Q2 if I don't receive the board ASAP
09:23 < dammsan> have you tried communicating by smoke signals? =)
09:23 < pinchartl> I tried calling them, they don't speak English
09:23 < pinchartl> and they don't react to my e-mails
09:24 < dammsan> pinchartl: i can give you H3 ES2 on Salvator-X, Salvator-XS and ULCB
09:24 < dammsan> via remote access
09:24 < pinchartl> dammsan: I can perform limited testing only through remote access :-/
09:24 < dammsan> i can hook up the VGA connector to VNC if that helps
09:24 < pinchartl> it's still useful
09:24 < pinchartl> but I need to test the LVDS output
09:24 < dammsan> not sure if ULCB has VGA
09:24 < pinchartl> that's the tricky one
09:24 < dammsan> yeah sorry no can do
09:25 < pinchartl> I know, not blaming you
09:25 < dammsan> i'll begin with Salvator-X H3 ES2
09:25 < pinchartl> I'm just saying that if the board doesn't arrive tomorrow, I won't deliver for Q2
09:25 < pinchartl> Friday is a public holiday here so they won't deliver anything
09:26 < pinchartl> and if I receive the board on Monday it will be too late for Q2
09:26 < pinchartl> anyway, that's it for me
09:26 < pinchartl> dammsan: your turn :-)
09:26 < dammsan> nothing to report from me really
09:27 < dammsan> i have review of your IPMMU DU series on my TODO still
09:27 < pinchartl> speaking of IPMMU, have you checked whether the IOMMU fixes for the v4.12-rc1 regressions have been merged ?
09:28 < dammsan> no sorry i did not
09:28 < geertu> pinchartl: They were merged last week
09:29 < geertu> through joro's tree, IIRC
09:29 < pinchartl> geertu: thank you !
09:29 < pinchartl> that's a relief
09:31 < pinchartl> next, Morimoto-san
09:31 < pinchartl> Since last meeting:
09:31 < pinchartl> - Posted remaining sound-related cleanup patches
09:31 < pinchartl> Almost all the patches have been accepted.
09:31 < pinchartl> - Cleanup the ALSA SoC framework
09:31 < pinchartl> This was Lars-Peter's plan, but nothing happen so far, so let's give it a try ourselves.
09:31 < pinchartl> For the next two weeks:
09:31 < pinchartl> - Continue ALSA SoC cleanup work
09:31 < pinchartl> Issues and Blockers:
09:31 < pinchartl> - HDMI legal issues
09:31 < pinchartl> Our Current Gen3 datasheet has HDMI related information, and because of it, we are upstreaming code. But it seems this is a compliance infraction. Now, our law team is checking the agreement with Synopsys and HDMI organization, and our current HDMI status, upstreamed code, ...
09:31 < pinchartl> While there is already upstreamed Synopsys HDMI code in Linux, this is a different issue. We have to keep distance from HDMI sound until this issue is cleared.
09:31 < pinchartl> the last point worries me
09:31 < pinchartl> Morimoto-san mentioned HDMI sound only
09:32 < pinchartl> I wonder what the implications are for HDMI video
09:33 < pinchartl> dammsan: do you have any information about that ?
09:33 < dammsan> no special information sorry
09:33 < dammsan> just smell wise it reminds me of SDHI
09:33 < dammsan> basically a lot of FUD
09:33 < dammsan> hopefully it will become less windy in the not so distant future
09:34 < dammsan> after people have sorted out the difference of documentation license, software license and open source and what not
09:34 < pinchartl> on the video output side we're mostly refactoring existing code so I don't plan to stop working on HDMI output for now
09:35 < dammsan> sounds fine with me
09:35 < dammsan> we will let you know if some special precaution is needed
09:35 < pinchartl> thank you
09:35 < pinchartl> next, Niklas
09:35 < pinchartl> Since last meeting:
09:35 < pinchartl> - [PATCH v3 0/2] v4l2-async: add subnotifier registration for subdevices
09:35 < pinchartl> - [PATCH v4 0/2] media: entity: add operation to help map DT node to media pad
09:35 < pinchartl> (merged to media-tree)
09:36 < pinchartl> For the next two weeks:
09:36 < pinchartl> - Try to address Hans comments on 'add subnotifier registration for subdevices'. Not sure how do make it work nice with v4l2, Laurent if possible I like to discuss options with you.
09:36 < pinchartl> - Look at the video device life time issues for VIN to for the Gen3 patch set.
09:36 < pinchartl> Issues and Blockers:
09:36 < pinchartl> - Not sire how to make v4l2
09:36 < pinchartl> I assume s/sire/sure/
09:36 < pinchartl> but I'm still not sire to understand that comment
09:37 < pinchartl> I suppose that's related to "add subnotifier registration for subdevices"
09:37 < neg> Yes :-) Sorry I'm not sure how to adress Hans comment about v4l2 incremental async, I have an idea but not sure it will work out
09:38 < pinchartl> we can discuss this later today if you're available
09:39 < neg> Plan is to try the idea and if it feels bad talk to you and seek guidance
09:40 < pinchartl> just ping me whenever convenient for you
09:41 < neg> thank you
09:41 < pinchartl> next, Ulrich
09:42 < uli___> so i've probed around the chromebook, and it is not very developer-friendly :)
09:42 < uli___> no uart, no jtag that i could find
09:42 < uli___> but you can reset the usb hub, and the data lines on the usb port change
09:42 < uli___> works reliably, and only requires a single write
09:42 < uli___> and no case opening
09:43 < uli___> also sent a serdev multiplexer prototype, to constructive feedback
09:43 < uli___> i'll incorporate that
09:43 < uli___> that also includes a max9260 i2c driver
09:43 < uli___> which works when talking to the max9260 itself
09:43 < uli___> but i cannot talk to anything behind the gmsl link
09:43 < dammsan> thanks for your efforts
09:43 < uli___> welcome
09:44 < dammsan> i'm interested in reproducing your chromebook modification
09:44 < uli___> you don't actually have to modify anything, that's the good part about it
09:44 < uli___> just hold a probe, or an led, or whatever you have to the usb data line
09:44 < uli___> and watch if the level changes
09:45 < dammsan> i'll see if i can incorporate that into my remote access setup somehow
09:45 < dammsan> together with the key press needed to boot
09:45 < uli___> robot arm? :)
09:45 < dammsan> hehe
09:46 < uli___> anyway, i need you to check if the max9259 are connected/powered properly
09:46 < uli___> because on this side of the link, everything looks fine
09:46 < uli___> all the registers are set up properly etc
09:47 < dammsan> yeah
09:48 < uli___> so that's it for me
09:48 < pinchartl> thank you
09:48 < pinchartl> Topic 2. Additional tasks for Q3/1
09:48 < dammsan> did you get schematics from public space?
09:48 < pinchartl> We have several candidates for additional tasks in Q3/1, including
09:48 < pinchartl> - Gen3 support rework for media controller
09:48 < pinchartl> - CSI2 virtual channel support
09:48 < pinchartl> - ADV7482 CSI-2 virtual channel configuration
09:48 < pinchartl> - V4L2 .s_stream() pad-aware operation
09:48 < pinchartl> - V4L2 multiplexed stream support (.s_stream(), frame descriptors)
09:48 < pinchartl> - RDACM20 driver upstream
09:48 < pinchartl> - MAX9286 driver upstream
09:48 < pinchartl> it also makes sense to continue with the Blanche-related tasks, and with the Chromebook tasks
09:48 < pinchartl> dammsan: I'd like your feedback on that though
09:49 < pinchartl> please all also feel free to propose other tasks that you believe would be useful
09:49 < dammsan> ok, on exactly what?
09:49 < pinchartl> dammsan: on Blanche & Chromebook
09:49 < dammsan> oh i see
09:49 < pinchartl> I don't know how much time/budget we can spend on that for Q3
09:49 < dammsan> yeah i guess they are currently blocked on me
09:50 < dammsan> we also may get access to V3H later on
09:50 < dammsan> but not for first batch most likely
09:51 < pinchartl> I want to send a proposal for Q3/1 additional tasks at the beginning of next week at the latest
09:51 < dammsan> uli: shall we try to unblock blanche later today?
09:51 < dammsan> please do, begin with task titles
09:52 < dammsan> so i can clear those with renesas side
09:53 < pinchartl> sounds good to me
09:53 < pinchartl> Topic 3. Next meeting
09:53 < neg> looks like it's my turn next at the hospital queue so I will drop of shortly. Will read backlog and comment on anything I think is interesting :-) Sorry for the scheduling conflict
09:53 < dammsan> i will be in the renesas offie tomorrow and thurday next week
09:54 < dammsan> so if task list updates can happen before that it would make review smooth
09:54 < pinchartl> as we still haven't decided on how to group multiple meetings in a single day, I propose scheduling the next meeting on Wednesday 2017-07-05 as usual
09:54 < pinchartl> and we can reschedule it if we can come to an agreement
09:54 < dammsan> sure
09:54 < pinchartl> dammsan: can I let you drive the meetings schedule discussion, as the request comes from you ?
09:55 < dammsan> sounds good
09:55 < pinchartl> thank you
09:55 < pinchartl> that's it for today
09:55 < jmondi> pinchartl: any update on v4l2 side?
09:56 < pinchartl> jmondi: not yet. I need to discuss it with Hans and Sakari, perhaps tomorrow or on Friday
09:56 < jmondi> and I admit I've not been following the remote setup issues, is 8 channel camera setup available now?
09:56 < jmondi> pinchartl: thanks
09:57 < pinchartl> jmondi: I think it's on port 8010 now
09:57 < pinchartl> dammsan: can you confirm ?
09:57 < dammsan> i have not reinstalled any boards
09:57 < uli___> dammsan: today i'm scheduled already, how about tomorrow?
09:57 < dammsan> uli___: sure, at what time?
09:58 < pinchartl> dammsan: could you reinstall the 8x camera boards too ?
09:58 < uli___> 10cest?
09:58 < uli___> i.e., in exactly 24 hours :)
09:58 < dammsan> uli___: my only appointment is between 16:00 and 17:00 JST
09:58 < jmondi> got "Permission denied (publickey,keyboard-interactive)." on port 8010
09:59 < dammsan> ok, 17:00 JST it is =)
09:59 < uli___> ok
09:59 < dammsan> jmondi: all gen3 boards have been disconnected
09:59 < dammsan> i'll begin by installing some H3 boards =)
10:00 < dammsan> then 8 x camera tomorrow if i can escape from the renesas office at some point =)
10:00 < pinchartl> I propose adjourning the meeting, does anyone second ?
10:00 < dammsan> sounds good
10:01 < dammsan> cya later guys
10:01 < pinchartl> meeting adjourned
10:01 < pinchartl> thank you all for attending, and have a nice day