From 55e3b2f45880faaf06f3c660ca9e8a6d9aa14bce Mon Sep 17 00:00:00 2001 From: Kuninori Morimoto Date: Mon, 9 Dec 2019 15:29:52 +0900 Subject: wiki: Porting wiki: Porting Chat Log Signed-off-by: Kuninori Morimoto --- wiki/Chat_log/20180726-mm-chatlog | 290 ++++++++++++++++++++++++++++++++++++++ 1 file changed, 290 insertions(+) create mode 100644 wiki/Chat_log/20180726-mm-chatlog (limited to 'wiki/Chat_log/20180726-mm-chatlog') diff --git a/wiki/Chat_log/20180726-mm-chatlog b/wiki/Chat_log/20180726-mm-chatlog new file mode 100644 index 0000000..a156dd6 --- /dev/null +++ b/wiki/Chat_log/20180726-mm-chatlog @@ -0,0 +1,290 @@ +Multimedia-chat-meeting-2018-07-26 + +10:15 < pinchartl> welcome to the multimedia meeting +10:15 < pinchartl> let's start with status updates +10:15 < pinchartl> * Jacopo +10:15 < pinchartl> Since last meeting: +10:15 < pinchartl> - soc camera removal +10:15 < pinchartl> - BSP drm/du patch fishing and analysis +10:15 < pinchartl> - v2 mt9v111 +10:15 < pinchartl> - Patch review: ov5640, a few bits of sccb, & other multimedia related stuff +10:15 < pinchartl> Until next meeting: +10:15 < pinchartl> - More du/drm BSP patch fishing and upporting +10:15 < pinchartl> - Possibly resume drm_bridge work (Laurent: priority of this task?) +10:15 < pinchartl> - Ebisu VIN+DU: is anyone working on it? Is it of interest? +10:15 < pinchartl> Issues and Blockers: +10:15 < pinchartl> - BSP patch handling is time consuming with not a lot of visible output. +10:15 < pinchartl> Is this known and OK ? +10:16 < pinchartl> to answer your last question, I think it's known and OK, as one of the metrics used to evaluate us is the number of patches "upported" +10:16 < pinchartl> (as in dropped from the BSP as a result from our work) +10:16 < pinchartl> handling BSP patches improves that metric, so it should be fine +10:16 < pinchartl> morimoto: any comment on that ? +10:17 < pinchartl> jmondi: any other comment ? +10:17 < jmondi> pinchartl: I see... I'm only concerned that the output might be not much compared to the time spent on that... +10:17 < jmondi> but I understand... +10:18 < jmondi> pinchartl: what about ebisu and drm_bridge format? what are priorities? +10:19 < pinchartl> I'll work on DU for Ebisu as part of an additional task related to DU LVDS PLL +10:20 < pinchartl> for VIN, yes, it's of interest +10:20 < jmondi> neg: are you on this? ^ +10:20 < pinchartl> and regarding drm_bridge format, it's still useful, but has low priority at the moment +10:21 < jmondi> pinchartl: I see... ok then, please proceed with status updates +10:21 < pinchartl> thanks +10:22 < pinchartl> one question though +10:22 < pinchartl> how far are we from dropping soc-camera ? +10:22 < jmondi> good question +10:22 < jmondi> on our side (ceu and sh) I guess we're done +10:23 < neg> jmondi: ? +10:23 < pinchartl> have we addressed all the SH boards ? +10:23 < jmondi> Hans is to clean up the ARM part, I asked if he needs help and he said he'll handle this +10:23 < jmondi> pinchartl: yes, the only remaning bits is defconfigs, which Mauro is bouncing to SH people +10:24 < pinchartl> great news, thanks +10:24 < jmondi> neg: Enabling VIN on Ebisu: do you have plans? +10:25 < neg> jmondi: no plan as I no one steped up to test it, do have plans for V3H as Sergie steped up as a tester +10:25 < pinchartl> jmondi: you can go for it then +10:26 < pinchartl> * Kieran +10:26 < pinchartl> Since last meeting: +10:26 < pinchartl> - PeriSchedule - Is this useful to anyone else? +10:26 < pinchartl> https://gitlab.com/PeriPeri/peri-schedule # +10:26 < pinchartl> https://periperi.gitlab.io/peri-schedule/ # +10:26 < pinchartl> https://periperi.gitlab.io/peri-schedule/periperi.ics # calendar feed> +10:26 < pinchartl> (Note, because we can't host arbitrary code/projects on OSDR Kieran has created a private group "periperi" at gitlab, if you want to host code +10:26 < pinchartl> there, just create an account and join the group, or ask him to add you +10:26 < pinchartl> to it) +10:26 < pinchartl> - DU/Interlaced respin - v5 posted +10:26 < pinchartl> - D3 Interlaced investigation +10:26 < pinchartl> - Reviewed driver for Aptina MT9V111 +10:26 < pinchartl> - Looked at PeriJect +10:26 < pinchartl> - Planning tasks with Jacopo +10:26 < pinchartl> - Created git-hooks for periupport +10:26 < pinchartl> - Rebased vsp1 writeback prototype +10:26 < pinchartl> Until next meeting: +10:26 < pinchartl> - Retest/examine partition algorithm limitation patch +10:26 < pinchartl> - Repost writeback prototype +10:26 < pinchartl> - Submit GMSL for Renesas-drivers? +10:26 < pinchartl> Issues and blockers: None +10:26 < jmondi> I need a tester or handle it from remote somehow +10:26 < pinchartl> kbingham: any comment ? +10:26 < pinchartl> jmondi: we should receive Ebisu boards soon(ish?), so I think you could find testers +10:26 < pinchartl> provided it doesn't require external hardware difficult to source +10:27 < kbingham> morimoto, I like your output (colorisation) on your peri-schedule example. ... +10:27 < kbingham> but I do worry - like pinchartl mentioned if one tool should do everything. ... though storing the data in a common place may not be terrible. +10:28 < kbingham> The part for me was linking to a generated ICS which is hosted so it integrates into my live google calendars, but we could automate the conversion of perischedule to ICS still and host somewhere. +10:28 < kbingham> It's mainly keeping the data live that matters ;) +10:29 < kbingham> Other than that ... no further comments from me at the moment. +10:29 < pinchartl> the concept of a task manager + bug tracker + calendar tool makes me wonder when we'll turn it into an e-mail client :-) +10:29 < kbingham> :D +10:29 < pinchartl> * Laurent +10:29 < pinchartl> Since last meeting: +10:29 < pinchartl> - Patch review +10:29 < pinchartl> - Additional tasks preparation +10:29 < pinchartl> - Part-time vacation +10:29 < pinchartl> Until next meeting: +10:29 < pinchartl> - Patch review +10:29 < pinchartl> - Additional tasks preparation +10:29 < pinchartl> - Revive the GMSL efforts +10:29 < pinchartl> Issues and blockers: None +10:29 < pinchartl> I'm back from holidays and fully operational now +10:29 < pinchartl> (well, as operational as possible at least +10:29 < pinchartl> ) +10:30 < jmondi> is that a good thing for us or not? +10:30 < pinchartl> any comment, anything I should look at with high priority ? +10:30 < pinchartl> jmondi: only you can tell +10:30 < jmondi> we managed to skip all the work while you were away +10:30 < kbingham> pinchartl, Will DU/Interlaced be able to make it to v4.19 ? +10:31 < kbingham> Oh ... actaully it may be blocked on me ... as I might need to find a way to limit on D3 still :( +10:31 < pinchartl> kbingham: for v4.19 I don't think so, we're too close to the merge window +10:31 < kbingham> Ok. +10:32 < kbingham> New target of v5 (v4.20) then :) +10:32 < pinchartl> that's reasonable I think +10:32 < pinchartl> * Morimoto-san +10:32 < pinchartl> Since last meeting: None +10:32 < pinchartl> Until next meeting: None +10:32 < pinchartl> Issues and Blockers: None +10:32 < pinchartl> morimoto: any comment? :-) +10:33 < pinchartl> * Niklas +10:33 < pinchartl> Since last meeting: +10:33 < pinchartl> - [PATCH 0/2] adv7180: fix format and frame interval +10:33 < pinchartl> - Talking to Hans about patchwork legacy. +10:33 < pinchartl> - Working on v4l2 virtual channel patches. +10:33 < pinchartl> - Reviewed potential CSI-2 VC patch-sets. +10:33 < pinchartl> - Worked with Mauro on STD merger of v4l patches. +10:33 < pinchartl> Until next meeting: +10:33 < pinchartl> - Vacation but hope to post next VC patch series. +10:33 < pinchartl> - Errata from latest datasheet for VIN and CSI-2 patches. +10:33 < pinchartl> Issues and blockers: None +10:33 < pinchartl> neg: any comment? +10:33 < pinchartl> when are your holidays exactly ? +10:33 < neg> no additional comment +10:34 < neg> Next week i will holiday and work, the week after I will be offline +10:34 < pinchartl> thank you +10:35 < pinchartl> * Simon (Kaneko-san) +10:35 < pinchartl> Since last meeting: +10:35 < pinchartl> - v1 and v2 of M3-N Audio support +10:35 < pinchartl> Until next meeting: +10:35 < pinchartl> - Testing M3-N Audio support +10:35 < pinchartl> - D3 Audio support +10:35 < pinchartl> Issues and blockers: +10:35 < pinchartl> - Some trouble finding patch to add Audio DMA node for R-Car D3 +10:35 * morimoto it is too fast, there is no chance to explain for me... +10:35 < pinchartl> morimoto: take your time :-) +10:35 < pinchartl> we won't ignore your messages +10:36 < pinchartl> Simon is gone... +10:36 < pinchartl> * Ulrich +10:36 < pinchartl> Since last meeting: +10:36 < pinchartl> - Reviewed Condor DU/LVDS/HDMI patch +10:36 < pinchartl> Until next meeting: None +10:36 < pinchartl> Issues and blockers: None +10:36 < pinchartl> uli___: any comment? +10:36 < uli___> none +10:36 < pinchartl> thanks +10:36 < pinchartl> morimoto: now you have our full attention :-) +10:37 < morimoto> ahh, thanks +10:38 < morimoto> but, there is still no rule for us how to handle periject +10:38 < morimoto> thus, it seems bike-shade story for me, for now +10:38 < pinchartl> it's still an important matter +10:39 < pinchartl> I've replied to most of the e-mails, I'll handle the remaining one today and tomorrow +10:39 < morimoto> I still don't understand why you guys doesn't like 1 tool for multi topics +10:39 < kbingham> morimoto, It seems to me that we should ... just start using it ? +10:39 < pinchartl> in my opinion, we're really moving in the right direction +10:39 < pinchartl> and I think we should be able to start using it soon +10:39 < morimoto> kbingham: I agree. maybe we are misunderstanding each +10:40 < pinchartl> thanks for pushing it to a git repo, that will help tracking changes +10:40 < morimoto> np +10:40 < kbingham> (in regards to the peri-schedule, I saw it as a specialised tool that would turn human readable text into a google calendar feed, and didn't necesarily expect that to be part of periject) +10:40 < pinchartl> regarding one tool for multiple purposes, I'm fine with that when the purposes are related +10:40 < kbingham> That code 'could' live in periject too. +10:41 < morimoto> kbingham: about google calendar, can you share it to us ? +10:41 < kbingham> morimoto, I have :) +10:41 < morimoto> I think google calendar is such feature +10:41 < morimoto> nice +10:41 < kbingham> morimoto, https://periperi.gitlab.io/peri-schedule/periperi.ics +10:42 < kbingham> It's generated by code which is currently hosted privatly on gitlab, (because it contains our group confidential information about our meetings, so I didn't want that to be unnecessarily public) +10:42 < morimoto> no, I think you can select me on google calendar +10:42 < pinchartl> regarding periject, I'd like to concentrate on tasks and BSP commits (bug reports) first as that's the main target +10:42 < kbingham> The ICS file can be added as a calendar feed in google-calendar or any other calendar tool +10:42 < pinchartl> once done with that, let's see what else would make sense to integrate in the tool +10:43 < kbingham> (sorry - not file - URL ... you add the URL - and google reads it when it's updated) +10:43 < kbingham> you don't need to download the file, and then uplaod or anything like that - it's a one add job :) +10:44 < kbingham> To do the same from periject would require a public url space ... that's all. +10:44 < pinchartl> something that I think would be useful is to start populating the repository with a set of real tasks (let's say about 10) +10:44 < pinchartl> in order to better understand how it works +10:44 < kbingham> ^ I agree :) +10:45 < kbingham> Is it one file per task ? +10:45 < morimoto> kbingham: you can select +10:45 < pinchartl> I don't like having multiple options +10:45 < morimoto> 1 file / 1 task or 1 file / multi tasks +10:45 < pinchartl> as in one task per file, or multiple tasks per filie +10:46 < pinchartl> or with or without white spaces to indent lines within a file +10:46 < pinchartl> we should document the file format clearly +10:46 < pinchartl> and stick to one format +10:46 < pinchartl> otherwise it will become messy with everybody having a slightly different style +10:46 < kbingham> peri-lint.py ;) +10:47 < pinchartl> kbingham: we need a git commit hook that will validate the format +10:47 < kbingham> pinchartl, morimoto-san already has a git hook in place to do sanity checks +10:47 < kbingham> So I'm sure that can be extended as necessary +10:48 < kbingham> And actually - I like that the code lives in the same repository as the data in this instance - because it will enforce everyone being on the latest version of the code - if they have the latest version of the data ;) +10:48 < pinchartl> I think we should continue this discussion over e-mail to ensure we keep track of it +10:48 < kbingham> (normally I would have separated) +10:48 < morimoto> about alignment, creating "tool rule" is OK for me. But 1 task / 1 file vs multi task / 1 file is "operation rule" for me +10:48 < pinchartl> kbingham: I have mixed feelings about that +10:48 < pinchartl> it makes the history pretty messy +10:49 < morimoto> I want to have flexibility +10:49 < pinchartl> morimoto: why is it an "operation rule" ? (and what do you mean by that exactly ?) +10:51 < morimoto> I understand "you" don't like it. But "otherone" might like it. I don't want to force to 1 opinion +10:51 < morimoto> If you don't like it, you just don't do it. +10:52 < pinchartl> it then becomes a real mess +10:52 < morimoto> This is my "operation rule". If tool forced it it is "tool rule" +10:52 < pinchartl> what if I'd prefer a third format and implement support for it ? will it be accepted as I won't force anyone to use it ? +10:52 < pinchartl> or a fourth, fifth format ? +10:52 < pinchartl> also, it's not true that nobody would be forced to do it +10:52 < pinchartl> as we'll all have to modify tasks not created by us +10:52 < pinchartl> or at least read them +10:53 < pinchartl> readability is very important, and comes from consistency +10:53 < pinchartl> same as for coding style, if we had no coding style rule in the linux kernel, it would be awful +10:53 < morimoto> Having coding rule is OK for me. +10:54 < morimoto> About task vs file, let's use 1st. I think you are misunderstanding +10:54 < morimoto> or I'm misunderstanding +10:55 < pinchartl> the reason why I think we should have a single task per file is that looking at the contents of the tasks directory should give us an overview of all tasks, with file names containing the task subjects +10:55 < pinchartl> you can't do that with multiple tasks per fil +10:55 < pinchartl> e +10:56 < pinchartl> but again, I think we should continue this discussion over e-mail, it's getting out of scope for multimedia +10:56 < morimoto> agree +10:57 < pinchartl> thank you +10:57 < pinchartl> Topic 2. Additional Multimedia Tasks for Q3 +10:57 < pinchartl> we have agreed on additional tasks last time, I'll submit task descriptions to Magnus before the end of the week +10:57 < neg> thanks for doing that +10:57 < pinchartl> kbingham: jmondi: have you sorted out how to split the work between the two of you ? +10:57 < pinchartl> neg: you're welcome +10:58 < jmondi> pinchartl: so far we've been working in going through the BSP patches realted to du/drm together +10:59 < pinchartl> do you need any help ? +10:59 < jmondi> so short answer is "not yet" +10:59 < kbingham> pinchartl, not a direct split, because we don't have a clear boundary +10:59 < pinchartl> ok :-) +10:59 < jmondi> Help yes, but I think going through email is easier +11:00 < pinchartl> fine with me +11:00 < pinchartl> Topic 3. Periperi Meetings +11:00 < pinchartl> The plan for an annual meeting in Paris around Kernel Recipes has been canceled, and the meeting will take place in Edinburgh around ELCE. +11:00 < pinchartl> I would still like to organize a multimedia code camp in September, the two candidate locations being Paris around Kernel Recipes and A Coruña around XDC. +11:00 < pinchartl> who would join such a code camp ? +11:00 < jmondi> if it is fine to send patches to periupport as we've been doing recently... we can identify patches to look into and have them reviewed commented by you and the team +11:00 < pinchartl> jmondi: yes, it's totally fine +11:00 < jmondi> sorry, I was still on previous point +11:01 < jmondi> pinchartl: o/ +11:01 < kbingham> pinchartl, I'm more likely to head to paris than A Coruna. +11:01 < jmondi> conference wide I would go for Paris, location wide for XDC +11:02 < jmondi> but I've never been to XDC, so it might be a pleasant surprise +11:02 < pinchartl> conference-wise I think XDC is more related to multimedia than KR +11:02 < pinchartl> also, with the very limited number of tickets for KR, we might not be able to attend the conference +11:02 < neg> would join both but as stated erlier I need to go back during weekend for the proposed dates. If we are in Paris I will rejoin if we are in A Coruna maybe not +11:03 < pinchartl> if we go to A Coruna should we then try to host the code camp before the conference ? +11:04 < jmondi> neg: shouldn't you be back on the weekend of the 22? +11:05 < jmondi> pinchartl: before the conference it would be 24th to 26th, after it would span to the 28th to the Monday after? +11:05 < jmondi> more or less +11:05 < neg> jmondi: yes the 22 I need to be in Stockholm. Other dates I'm free +11:06 < jmondi> neg: both conferences are 26th to 28th +11:06 < pinchartl> jmondi: the conference is on 26-28 +11:06 < pinchartl> neg: sorry I got the dates wrong +11:06 < jmondi> Embedded recepies is 24th to 26th +11:06 < pinchartl> we should then have the code camp after the conference +11:06 < jmondi> pinchartl: I think so +11:07 < pinchartl> regardless of whether it's in Paris or Spain +11:07 < pinchartl> I think we decided for the week before because Wolfram wasn't available the week after (if I recall correctly) +11:07 < pinchartl> we if it's just multimedia, it doesn't matter +11:07 < pinchartl> s/we/but/ +11:07 < neg> if the reschdeuling of code camp to after the conference is possible it would be benificial for me and then I would vote for Spain :-) +11:09 < pinchartl> ok +11:09 < pinchartl> I'll sleep over that, let's discuss it again in the next few days +11:09 < pinchartl> Topic 4. Next Meeting +11:09 < pinchartl> that's settled already, 2 weeks from now +11:09 < pinchartl> any other topic to discuss ? +11:10 < neg> not from me +11:10 < pinchartl> anyone else ? +11:10 < pinchartl> jmondi: ? +11:10 < pinchartl> kbingham: ? +11:10 < pinchartl> morimoto: ? +11:10 < pinchartl> uli___: ? +11:10 < kbingham[m]> Good here +11:10 < morimoto> not from me +11:10 < uli___> nope +11:11 < pinchartl> in that case I declare the meeting adjourned +11:11 < pinchartl> thank you all for attending +11:11 < pinchartl> and have a nice day +11:11 < jmondi> bye bye, thanks +11:11 < pinchartl> oh, a generic question +11:11 < pinchartl> shimoda: if you're still here +11:11 < neg> thanks all +11:11 < pinchartl> when do you plan to ship the Ebisu boards ? +11:11 < shimoda> pinchartl: yes? +11:12 < shimoda> ah, morimoto-san is doing export control +11:12 < pinchartl> morimoto: the question is for you then :-) +11:12 < morimoto> pinchartl: now it is under export paper work +11:12 < pinchartl> do you have an idea when it would ship ? my additional task due for 9/M will depend on access to an Ebisu board +11:13 < morimoto> I think paper work will takes 1 week or 2week +11:13 < morimoto> but no +11:13 < morimoto> but no guarantee +11:13 < pinchartl> ok. if it gets delayed, could you please keep me informed ? +11:14 < morimoto> OK +11:14 < pinchartl> thank you very much +11:14 < pinchartl> no other question :-) +11:14 < pinchartl> have a nice day everybody -- cgit v1.2.3