path: root/wiki/Chat_log/20180809-core-chatlog

Core-chat-meeting-2018-08-09

09:31 < geertu> Welcome to today's Core Group Chat!
09:31 < geertu> Agenda:
09:31 < geertu> 1. Status Updates
09:31 < geertu> 2. Discussion Topics
09:31 < geertu> Topic 1. Status updates
09:31 < geertu> nA) What have we done since last time:
09:31 < geertu> Marek worked on U-Boot (USB VBUS and PHY on Gen2, HS400 on Gen3) and Linux
09:31 < geertu> (PCIe L0s/L1 handling on Gen3).
09:31 < geertu> Morimoto-san worked on PeriJect and Ebisu-4D export.
09:31 < geertu> Shimoda-san submitted PWM patches, and provided LTSI v4.14 snapshot
09:31 < geertu> feedback.
09:31 < geertu> Geert submmited v2 of R-Car gen3 OSC and RCLK improvements, reviewed lots
09:31 < geertu> of patches, enabled the Global Timer on Cortex-A9 MPCore SoCs,
09:31 < geertu> sub-maintained LSTI, and started looking into QEMU GPIO handling.
09:32 < geertu> B) What we plan to do till next time:
09:32 < geertu> Kaneko-san will test the M3-N CPUFreq upport.
09:32 < geertu> Marek will continue U-Boot SDHI HS400 support.
09:32 < geertu> Morimoto-san will ship Ebisu-4D boards.
09:32 < geertu> Shimoda-san will check the new LTSI v4.14 snapshot test results from the
09:32 < geertu> RVC test team, and will pave the way forward for IPMMU.
09:32 < geertu> Geert will review LTSI submissions during the merge window.
09:33 < geertu> C) Problems we have currently:
09:33 < geertu> Linus postponed v4.18 by one week, conflicting with Geert's holiday
09:33 < geertu> schedule.  Hence next renesas-drivers release will be
09:33 < geertu> renesas-drivers-2018-08-21-v4.18.
09:33 < geertu> Anything I missed?
09:34 < geertu> s/submmited/submitted/
09:35 < geertu> Lazy Summer (although it's a lot colder again)...
09:35 < Marex> geertu: well, HS400 in U-Boot is dragging on because the MMC maintainer is horribly slow to respond ... I might need to bypass him again at some point
09:35 < geertu> Marex: I'm sure you'll handle that fine
09:35 < horms> ... and for the same reasons my LTSI submission will be up to v4.18-rc8 rather than v4.18
09:36 < Marex> geertu: just like last time ... sadly ... we might be looking for a new mmc maintainer ;-)
09:36 < geertu> horms: Np, as v4.18 will (should) be released long before the LTSI merge window closes
09:36 < horms> not before the end of this week which is when I will be sending the next round of patches
09:36 < horms> before my holiday next week
09:37 < horms> otherwise it all falls to the last week of the merge window and the wheels could easily fall off at that point
09:37 < horms> I will follow-up with any patches added between v4.18-rc8 and v4.18
09:37 < geertu> horms: Postponing to the last week makes it more difficult for Intel to try to counter Renesas again ;-)
09:38 < horms> That is true
09:38 < geertu> (although they already know our plan anyway)
09:38 < geertu> Topic 2. Discussion Topics
09:38 < horms> But I'd rather de-risk other parts aspects of the project
09:39 < geertu> horms: Sure, thx a lot!
09:39 < geertu> morimoto: periject?
09:39 < wsa> what's with that Intel vs Renesas talk?
09:40 < horms> wsa: its a story from many years ago
09:40 < morimoto> geertu: sorry I didn't read text, but periject what ?
09:40 < horms> wsa: whereby there were two major contributors to LTSI. Renesas and Intel. And magically Intel had slightly more patches than Renesas
09:40 < geertu> wsa: Intel ended up backporting ca. 3 more patches than Renesas
09:41 < geertu> morimoto: Can we discuss periject?
09:41 < wsa> so they could claim #1
09:41 < wsa> ?
09:41 < horms> They also caused a hidious conflict with our work
09:41 < horms> wsa: yes, that is the theory
09:42 < morimoto> geertu: yes
09:42 < wsa> horms: bastards, I will only buy AMD cpus from now on!
09:42 < wsa> ;)
09:42 < horms> wsa: I got over it in time
09:43 < wsa> "There is no politics in open source"
09:44 < geertu> morimoto: The mic is yours!
09:44 < morimoto> thanks.
09:45 < morimoto> As you know I already posted periject on gitlab. you can try it.
09:45 < morimoto> I think "tool feature" is not yet 100% but almost OK
09:45 < geertu> morimoto: You have rebased the master branch?
09:45 < morimoto> Ah... yes
09:46 < morimoto> sorry, it is still v0.x version
09:46 < morimoto> After v1.0, I will not
09:46 < pinchartl> if I may chime in on this topic (being the main source of controversy...)
09:46 < pinchartl> I think we still haven't agreed on the development process
09:46 < pinchartl> and I don't see how we can build a tool to support a process if we don't define the process first
09:47 < pinchartl> that's been my issue since the beginning
09:47 < pinchartl> the face to face discussion Morimoto-san and I had in Tokyo answered lots of my questions
09:47 < pinchartl> and I tried to summarize the process discussions on the periperi mailing list
09:47 < pinchartl> but the mail thread quickly died
09:47 < pinchartl> to word this differently, I think we need to agree about what we want to do before doing it
09:48 < morimoto> 1 question. what is your problem ??
09:49 < wsa> what are the open questions there? I am fine with trying morimoto-san's tools for now...
09:49 < damm> i think we should split the process discussion from the tool
09:49 < damm> that said, i agree that process is important
09:49 < pinchartl> morimoto: my issue is that we still don't know what we want to do
09:49 < pinchartl> there have been lots of bikeshedding discussions about the tool
09:50 < pinchartl> about the format of stored data
09:50 < morimoto> OK, yes, let's split the topic
09:50 < pinchartl> and other topics
09:50 < pinchartl> and to answer those questions, we first need to know what we want to do
09:50 < pinchartl> at the moment, I see a tool, and I have no idea how to use it
09:50 < damm> i don't disagree
09:50 < pinchartl> for instance
09:50 < pinchartl> for your last report
09:50 < pinchartl> you tried to use the tool to generate the e-mail
09:51 < pinchartl> there were no B) and C)
09:51 < pinchartl> which isn't surprising
09:51 < wsa> we are at the tool level again
09:51 < pinchartl> as the tool doesn't support tracking future work plans at a bi-weekly level
09:51 < pinchartl> nor does it support tracking blockers
09:51 < wsa> I agree that "reporting emails" should be maybe a second or third step for the tool
09:51 < pinchartl> so, as an experiment, you tried adding that information to the task description
09:51 < pinchartl> it then ended up in the generated report
09:52 < pinchartl> but that's a hack
09:52 < damm> but does it blend?
09:52 < pinchartl> and if we start using a tool without knowing what we want to do it with
09:52 < pinchartl> we'll keep making hacks like that
09:52 < pinchartl> in different ways, for different people
09:52 < pinchartl> it will quickly become unusable
09:52 < wsa> we first need to deal with handling tasks, or?
09:52 < pinchartl> personally speaking, the logical order would be
09:53 < pinchartl> 1. what are the issues we have?
09:53 < pinchartl> 2. how do we want to solve them?
09:53 < pinchartl> 3. how do we implement tools to support that?
09:53 < wsa> and I would still like to understand why we can't use the tool for that right now...
09:53 < pinchartl> if we start by 3, then 2, then 1, I don't see how it could work :-)
09:53 < morimoto> wsa: +1
09:54 < morimoto> Many times I explain it ...
09:54 < wsa> and I think it is not only about the issues we have but also about the issue Morimoto-san is having
09:54 < wsa> issues
09:54 < pinchartl> wsa: we have a mostly free-formed text format as a database. that won't work, we will end up with one format per person depending on personal preferences
09:54 < damm> pinchartl: i thought that it was the germans that liked process and order?
09:54 < pinchartl> see the discussion of 1 task per file vs. several tasks per file
09:55 < pinchartl> or Morimoto-san's today's e-mail report with free-formed B) and C) in the task description
09:55 < pinchartl> we haven't really started using this, and it's already forking :-)
09:56 < damm> my opinion is that people should use the tools that they like
09:56 < wsa> I do like processes, but there is just talk :)
09:56 < pinchartl> at the moment I feel that we're being given a database engine and we're told it's our bug and task tracker
09:56 < damm> but some interface/format is needed together with a known process
09:57 < pinchartl> what I'd like to get is jira/bugzilla/whatever. not those tools in particular, but something that operates at a similar level. with a defined work flow
09:57 < damm> how about the people that like processes come up with a process proposal?
09:58 < damm> i am for sure in the process camp
09:58 < damm> the we have a tool camp as well
09:58 < pinchartl> damm: let me locate the e-mail thread
09:58 < damm> once both are in semi-ok order we merge
09:58 < morimoto> I explained many times, but, I can create "Tool", but, you can create "Rule"
09:58 < damm> how many people are in the pro-process discussion camp?
09:59 < kbingham> I'm in the ... I want something that works camp ...
09:59 < damm> obviously not morimoto-san
09:59 < wsa> okay, let me summarize my take:
09:59 < pinchartl> morimoto: but my point is that the tool should be based on the rules, not the other way around
09:59 < damm> obviously laurent likes process
09:59 < kbingham> And ... I'm confused why we're writting our own tools when opensource tools already exist :)
09:59 < morimoto> pinchartl: yes, and, no my opinion
10:00 < morimoto> My opinion is like this
10:00 < damm> kibingham: i guess we need to know which problem we are solving first
10:00 < wsa> even the current process works for me as groupleader. not perfect, but works. I see Morimoto-san has problems reporting our progress upwards. I'd surely like to have this issue solved, it is important for all of us.
10:01 < morimoto> You can use Emacs, or VIM. it is editor. but, what you can write is under your rule.
10:01 < pinchartl> "Re: [periperi] Peri Tool Next" from "Date: Sat, 14 Jul 2018 05:27:30 +0300"
10:01 < pinchartl> Message-ID: <2412691.0qdXsgRgLR@avalon>
10:01 < wsa> I don't think I need details on this, because I'd think there is a lot of internal information and culture involved.
10:01 < damm> so why can't we simply do a work split?
10:01 < damm> break out the process discussion into a group that cares?
10:02 < damm> and figure out what we _need_ to do
10:02 < pinchartl> damm: that would work for me
10:02 < damm> then we can apply that to tools or whatnot
10:02 < damm> so currently the process group includes pinchartl and myself
10:03 < morimoto> my opinion is, try 1st, fix 2nd, rule 3rd :)
10:03 < damm> morimoto: that's why you are not in the process group =)
10:03 < morimoto> ok :)
10:03 < damm> morimoto: you are however very welcome to join if you'd like
10:03 < morimoto> I still don't understand what is the issue...
10:04 < damm> morimoto: you and i have discussed this before
10:04 < pinchartl> morimoto: ok, to make it very simple, my first issue is that with today's version of the tool, file format and documentation, I have no idea how to use it
10:04 < damm> if you get interested in the process let me know
10:05 < damm> pinchartl: you and i are at the same wave length
10:05 < morimoto> pinchartl: yeah, I don't know too. thus, it is "trial"
10:05 < pinchartl> damm: I don't know if that's a good or bad thing :)
10:05 < damm> but lets give morimoto-san freedom to poke around
10:05 < pinchartl> morimoto: shouldn't the trial phase be done separately from the production phase ? :-)
10:06 < damm> you are using process-think now =)
10:06 < damm> lets take that elsewhere
10:06 < morimoto> My opinion is,
10:06 < wsa> damm: I assume you know what is needed from the Renesas side for that process?
10:07 < damm> sort of
10:07 < morimoto> we can use tool to get all information. But we don't know how many information we have or we need, so far
10:07 < morimoto> So far I know is
10:07 < morimoto> 1) task, 2) BSP list
10:08 < morimoto> For each purpose, we can use 1) task/mm for example, 2) task/bsp, etc
10:08 < morimoto> each folder has each rule
10:08 < morimoto> few rule is it should use "clear Title", etc
10:09 < morimoto> we can create each rule for each purpose / each folder
10:09 < morimoto> Because of this, "tool" have limited rule
10:10 < morimoto> If you want to have "fixed" fule/format, you can create it
10:10 < morimoto> and force it by git hook ?
10:10 < morimoto> it is operation rule, not tool rule
10:10 < morimoto> this is my basic idea
10:11 < pinchartl> morimoto: I think the tool should enforce operation rules. otherwise they won't be enforced, and the "database" will be "corrupted" all the time
10:11 < damm> morimoto: this discussion is similar to someone wanting to cook a pancake and asks how to do it, but the answer is a swiss army knife. =)
10:12 < morimoto> pinchartl: tool has operation rules. like "status:" tag
10:12 < wsa> how urgent is all this?
10:12 < morimoto> ? sorry, what does your "operation rule" ?
10:12 < wsa> I am under the impression Morimoto-san may need the tool to create proper reports so we can be evaluated correctly
10:12 < wsa> but i may be wrong
10:13 < damm> wsa: i have my own tools to monitor upstream contribution rate for each group
10:13 < pinchartl> wsa: I think you're right, but it's not just the tool that's needed for that, it's also the data. reports can't be created before we all start using the tool and keep the data up to date
10:13 < morimoto> wsa: creating reports is option, not purpose
10:13 < wsa> ok
10:13 < wsa> that relieves me
10:13 < damm> and to avoid spending time unwisely i don't bother you with that unless things are not progressing as expected
10:14 < wsa> I have absolutely zero doubts in trusting Magnus and Laurent figuring out an awesome process...
10:14 < damm> it would be interesting to hear how shimoda-san and morimoto-san would like to receive information how to handle those BSP upporting lists
10:15 < wsa> ... yet they are both probably the busiest people we have?
10:15 < damm> are we in a rush somehow?
10:16 < wsa> that was my initial question. i thought so and I am happy I was wrong :)
10:16 < pinchartl> wsa: I think we're all busy. I agree however that I might be one of the people who sometimes has the hardest time setting priorities straight :)
10:16 < damm> pinchartl: it is good to see that someone cares
10:16 < damm> pinchartl: shall we take that process discussion elsewhere?
10:17 < pinchartl> and as we're all busy, if this new tools requires spending more time to keep data up to date, I'm pretty sure it won't happen. I don't see how it could be adopted unless it makes everybody's life easier
10:17 < pinchartl> damm: sure. on the mailing list ? or do you have another preference ?
10:18 < damm> pinchartl: lets use private email to schedule a video conference you and me
10:19 < pinchartl> is anyone else interested ?
10:20 < geertu> I'm interested in the outcome
10:20 < kbingham> perhaps in the outcome of discussions :) - and the list of features required
10:21 < pinchartl> alright
10:21 < pinchartl> damm: seems like we have a date then :-)
10:21 < damm> pinchartl: yay
10:22 < pinchartl> ok, let's do that
10:23 < wsa> damm: you have a second now to chat (IRC or Hangout)?
10:25 < pinchartl> damm: mail sent
10:26  * kbingham posts "PeriZilla" to "PeriPeri" and awaits to be laughed at 
10:27 < pinchartl> damm: seems like our e-mails have crossed each other :)
10:29 < geertu> kbingham: is there a perl-bugzilla, too?
10:29 < kbingham> geertu, If not you could write one :D
10:29  * geertu notices the similarities between peri and perl
10:30 < uli_> geertu: bugzilla itself is written in perl; i'd be surprised if there weren't
10:30 < geertu> Are we finished with core?
10:30 < kbingham> geertu, https://devzing.com/blog/index.php/access-bugzilla-from-perl/
10:30 < geertu> Any other topics to discuss?
10:34 < wsa> damm: you have a second now to chat (IRC or Hangout)?
10:34 < damm> wsa: hangouts pls
10:34 < pinchartl> geertu: not for me. I'm fine proceeding with multimedia
10:34 < geertu> OK.
10:36 < geertu> Thanks for joining, and have a nice continued day.