IRC logs for #cip for Thursday, 2020-01-16

*** lchlan has joined #cip00:20
*** rynofinn has joined #cip00:33
*** rynofinn has quit IRC00:42
*** dl9pf has quit IRC00:42
*** lchlan has quit IRC00:42
*** rynofinn has joined #cip00:59
*** punit has joined #cip01:13
*** punit_ has joined #cip01:29
*** punit has left #cip01:29
*** rynofinn has quit IRC01:31
*** punit_ has quit IRC01:32
*** punit has joined #cip02:10
*** punit has quit IRC02:11
*** punit has joined #cip02:12
*** punit has quit IRC02:17
*** punit has joined #cip02:17
*** punit has joined #cip03:04
*** punit has quit IRC03:07
*** punit has joined #cip03:07
*** punit has quit IRC03:10
*** punit has joined #cip03:11
*** dl9pf has joined #cip06:55
*** lchlan has joined #cip06:59
*** rajm has joined #cip07:09
*** rynofinn has joined #cip07:11
*** therisen has joined #cip07:39
*** masashi910 has joined #cip07:58
*** pave1 has joined #cip08:57
*** fujita3 has joined #cip08:59
masashi910#startmeeting CIP IRC weekly meeting09:00
brlogger`Meeting started Thu Jan 16 09:00:02 2020 UTC and is due to finish in 60 minutes.  The chair is masashi910. Information about MeetBot at http://wiki.debian.org/MeetBot.09:00
brlogger`Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.09:00
brlogger`The meeting name has been set to 'cip_irc_weekly_meeting'09:00
*** brlogger` changes topic to " (Meeting topic: CIP IRC weekly meeting)"09:00
masashi910#topic rollcall09:00
*** brlogger` changes topic to "rollcall (Meeting topic: CIP IRC weekly meeting)"09:00
pave1hi09:00
masashi910please say hi if you're around09:00
fujita3hi09:00
masashi910#topic AI review09:01
*** brlogger` changes topic to "AI review (Meeting topic: CIP IRC weekly meeting)"09:01
masashi9101. Combine rootfilesystem with kselftest binary - Iwamatsu-san09:01
masashi910== Quote from Iwamatsu-san ==09:01
masashi910I tested on LAVA. But not finished.09:01
masashi910  https://lava.ciplatform.org/scheduler/job/949909:01
masashi910====09:01
masashi910I keep this AI open.09:01
masashi9102. Document a process on how to add tests to the CIP test setup - patersonc09:02
masashi910patersonc: are you around?09:02
pave1It seems that there's just three of us here.09:02
masashi910pave1: right...09:02
patersoncI'm in traffic09:02
bwh(hi))09:02
masashi910patersonc: Hi!09:03
patersonc15 mins I'll be online09:03
masashi910bwh: Hi!09:03
patersoncSorry09:03
masashi910patersonc: I see. Then, I will revisit your topics later.09:03
masashi9103. Arrangement of F2F Kernel Meeting in Nuremberg - masashi91009:04
masashi910https://doodle.com/poll/73if9scwqrh5ge3809:04
*** kazu has joined #cip09:04
masashi910I keep this doodle poll open for a while, so, please fill your preference if you are interested in joining.09:04
masashi910If you would like to attend remotely, please also let me know.09:05
masashi9104. Add config for qemux86-64 and BBB to both 4.4 and 4.19 - Iwamatsu-san09:05
*** kazu71 has joined #cip09:05
*** kazu has quit IRC09:05
masashi910== Quote from Iwamatsu-san ==09:05
masashi910Qemux86-64's config already merged to cip-kernel-config.09:05
masashi910BBB config is under testing. Please keep this A.I.09:05
masashi910====09:05
masashi910So, I keep this AI open.09:06
masashi910#topic kernel maintenance updates09:06
*** brlogger` changes topic to "kernel maintenance updates (Meeting topic: CIP IRC weekly meeting)"09:06
masashi910Pavel-san, do you have updates?09:06
pave1I did reviews on 4.19.94, 95 and 96. New 4.19-cip-rt kernel was released.09:07
masashi910pave1: Thanks very much for your works!09:08
bwhI've backported fixes for several CVEs, which should go into stable updates soon09:09
masashi910bwh: Great! Thanks very much!09:09
wenshi, I'm late09:09
masashi910wens: Hi! Do you have updates?09:10
wensI did a run of classify-failed-patches, haven't posted the results yet09:10
wensdidn't see anything critical09:10
masashi910wens: Thanks very much for your works!09:11
pave1bwh: There was/is discussion on the mailing list regarding to 4.19 end-of-life09:11
pave1Do we have indications that someone else (Debian?) will maintain 4.19 after regular -stable maintainance ends?09:12
bwhNo but I may take that on for Debian if necessary09:13
wensIIRC buster is on 4.19?09:14
pave1Ok :-).09:14
bwhwens: yes09:14
masashi910pave1: BTW do you have pointers for such a discussion?09:14
pave1masashi910: It was your mail AFAICT, "Kernel Team Roadmap for CY2020", "1. Readiness for the self-maintenance09:15
masashi910pave1: I see. :)09:15
masashi910Any other topics?09:16
masashi910309:16
masashi910209:16
masashi910109:16
masashi910Because Chris-san is not here, I will jump to CIP Core session.09:17
masashi910#topic CIP Core09:17
*** brlogger` changes topic to "CIP Core (Meeting topic: CIP IRC weekly meeting)"09:17
kazu71hello09:17
masashi910kazu71: the floor is yours09:18
kazu71It's time for me, or chris?09:18
kazu71ah, chis not attend today sorry09:18
masashi910kazu71: it's yours09:18
kazu711. Update PDP to v3.1 https://lists.cip-project.org/pipermail/cip-dev/2020-January/004143.html09:18
kazu71Add a rule to satisfy all run-time dependencies09:18
kazu71cip-pkglist scripts also updated to check the deps automatically09:18
kazu712. Created the 2nd package proposal for Debian minimum base system https://lists.cip-project.org/pipermail/cip-dev/2020-January/004160.html09:19
kazu7158 source packages and 84 binary packages are inclulded09:19
kazu71I though that these packages should be discussed before (or in parallel with) the security WG proposal because there are many overwraped packages09:19
kazu71(To CIP Core member company) Please tell your opinion about this proposal by the next TSC :)09:19
kazu713. Generic profile: (WIP) Support cip-kenrel-config, testing isar-cip-core with linux-cip-ci09:19
kazu71That's all from me09:19
masashi910kazu71: thank you for your works!09:19
masashi910Any questions or comments?09:20
masashi910any other topics?09:20
fujita3 kazu71: it is better to review the security WG proposal in parallel09:20
kazu71fujita3 OK, I'm discussing with Kent about this topic already09:21
fujita3kazu71: shoud we check both source packages and binary packages?09:21
fujita3 source packages have higher priority?09:22
kazu71fujita3 Yes, but binary packages are the main target of this review09:22
kazu71binary packages are generated from the corresponding source package09:22
fujita3ok. binary is higher09:22
* patersonc arrives09:23
masashi910any other topics?09:23
masashi910309:23
masashi910209:23
masashi910109:23
masashi910Because Chris-san just joined, let me go back to the testing.09:23
kazu71patersonc Hello!09:23
masashi910#topic Kernel testing09:24
*** brlogger` changes topic to "Kernel testing (Meeting topic: CIP IRC weekly meeting)"09:24
masashi910patersonc: the floor is yours09:24
patersoncHello09:24
patersoncI started looking into RT testing09:24
patersoncPlease see my email to cip-dev.09:25
patersoncIt would be good to get some input.09:25
pave1Thanks for starting this! :-)09:25
patersoncMain discussion points are a) what tests we should run b) what packages to add to cip-core09:25
pave1So I guess even cyclictest alone is useful, but it would be nice to run it with some background load.09:26
patersoncpave1: Okay09:26
patersoncI played around with hackbench before09:26
patersoncIt would be good to know what/how the RT project does09:26
kazu71patersonc Thank you for the updates. a) depends on b), is it correct?09:26
kazu71Oops, b) depends on a) I meant...09:26
pave1I was using kernel build (make -j 5) as a background load..09:27
patersonckazu71: I guess so :)09:27
pave1Would that be possible/easy to do?09:27
patersoncpave1: That would mean including the toolchains in cip-core. Is this something already done kazu71 ?09:28
pave1Yep, so that does not sound exactly easy.09:28
patersoncOr we could add things like Dhrystone or other benchmarks09:28
patersoncLet me take an action to ask the RT project what they do09:28
pave1Dhrystone I'd say is not exactly suitable.09:29
patersoncWe may as well do the same so we can compare results easier09:29
pave1We may want to replicate what they do, yes; but if we discover something else that results in high latencies, we get bonus points.09:29
pave1Perphaps we can talk after the meeting?09:29
patersoncpave1: Sure09:30
masashi910patersonc: noted. you take the AI "ask RT project what they do for RT testing."09:30
masashi910patersonc: Thanks for your updates!09:30
patersoncpave1: Where was the discussion about extending v4.19LTS?09:30
patersoncIs Iwamatsu-san online?09:30
masashi910patersonc: no. he is not attending this call.09:31
patersoncOkay. Thanks09:31
bwhpatersonc: So far that was private mail09:31
pave1patersonc: "Kernel Team Roadmap for CY2020", I see you in the cc list.09:31
patersoncAh okay. I thought you meant a discussion on linux-stable09:31
masashi910patersonc: Yes. I sent the local mail to you.09:32
patersoncThanks09:32
masashi910patersonc: Thanks for your updates!09:32
patersoncNo updates on my action btw.09:32
* patersonc testing report done.09:32
masashi910patersonc: I see. Noted.09:32
masashi910Any other topics?09:32
masashi910309:33
masashi910209:33
masashi910109:33
masashi910#topic Software update09:33
*** brlogger` changes topic to "Software update (Meeting topic: CIP IRC weekly meeting)"09:33
masashi910Quote from Suzuki-san "SW Updates WG doesn't have any update this week."09:33
masashi910any other topics?09:33
masashi910309:33
masashi910209:33
masashi910109:33
masashi910#topic AOB09:33
*** brlogger` changes topic to "AOB (Meeting topic: CIP IRC weekly meeting)"09:33
masashi910Are there any business matters to discuss?09:34
patersoncHas anyone seen the email from Dennis Semakin?09:34
kazu71sorry, not yet09:34
masashi910patersonc: What about?09:34
pave1Seen: Obtain full set of source code for cip image09:34
patersoncMaybe we should add a wiki page describing how to customise CIP-Core09:35
pave1It is more of build system question.09:35
patersonc(email is on cip-dev btw)09:35
pave1patersonc: We should be able to simply pointer to someone else's documentation, no?09:35
patersoncpave1: Maybe, but the more information we share on such topics the more people who will be drawn to the project09:36
patersoncIt's good to see someone trying to use our Kernel/FS09:37
masashi910patersonc: Thanks for raising this issue. Will continue to discuss next week?09:38
patersoncSure.09:38
patersoncI've now booked hotel/flights for Embedded World now btw.09:39
masashi910patersonc: Great!09:39
patersoncJust for the two days of CIP meetings though09:39
masashi910same here.09:39
masashi910Any other business matters to discuss?09:40
masashi910309:40
masashi910209:40
masashi910109:40
masashi910#endmeeting09:40
brlogger`Meeting ended Thu Jan 16 09:40:33 2020 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)09:40
brlogger`Minutes:        https://irclogs.baserock.org/meetings/cip/2020/01/cip.2020-01-16-09.00.html09:40
brlogger`Minutes (text): https://irclogs.baserock.org/meetings/cip/2020/01/cip.2020-01-16-09.00.txt09:40
brlogger`Log:            https://irclogs.baserock.org/meetings/cip/2020/01/cip.2020-01-16-09.00.log.html09:40
*** brlogger` changes topic to "Civil Infrastructure Platform Project. Find the logs at https://irclogs.baserock.org/cip/"09:40
masashi910Thanks very much! See you nepave1xt week!09:40
pave1Thank you!09:40
wensthank you09:40
fujita3thank you09:41
patersoncThanks. Apologies for being late.09:42
kazu71patersonc "toolchain" above means... some software to make CPU/memory/disk loads that affects cyclictest results?09:42
pave1bwh: If you/Debian is willing to maintain 4.19-stable in 2021+, I believe easiest way forward for CIP project is to simply keep using that.09:42
patersonckazu71: I was referring to build-essential etc.09:43
pave1bwh: If not, I believe best way forward for CIP would be to take over 4.19-stable maintainance; but I believe that would be more effort.09:43
kazu71patersonc In the target rootfs? if so, isar-cip-core could provide them09:43
patersoncpavel: More effort, but wasn't this the original aim of the project? We should be doing it already for 4.4 but we got lucky :)09:44
kazu71patersonc Deby provides (cross-)toolchain but it's not for target but for development environment09:44
patersonckazu71: Sure. I guess we're going to have to have a 'kernel testing' version of cip-core.09:45
kazu71patersonc I think so too09:45
pave1patersonc: Well, our goal is really _subset_ of the kernel. If we took over -stable maintainence, we'd have to maintain _whole_ kernel.09:45
pave1patersonc: I still believe it would be good idea, because we'd get more testing if our tree was exposed as "stable".09:46
patersoncpave1: True. sorry. Whether Greg would let us take over the 'official' maintenance of 4.19-stable is another question ;)09:46
pave1patersonc: OTOH if Debian wants to do the job, I obviously would not complain.09:46
kazu71patersonc BTW, are there any requests about the profiles (tiny/generic) for kernel/RT testing now? Should we provide either or both profiles for kernel testing?09:47
pave1patersonc: I'm pretty sure Greg would let us do that if noone else is interested.09:47
patersonckazu71: I'll start drawing up a list. I know that Iwamatsu-san created his own RFS for LTP testing, presumably because there were missing packages in cip-core09:48
patersoncpave1: :)09:48
masashi910kazu71: please let me know your opinion about Dennis's mail and Chris-san's comment.09:48
masashi910kazu71: Have you read Dennis mail?09:49
kazu71patersonc OK, we will discuss about this topic F2F09:49
kazu71masashi910 Partially. Maybe CIP Core should reply with information Dennis is asking09:50
pave1patersonc: Some background load is important for cyclictest. Daniel provided loads he is using on the mailing list.09:51
masashi910kazu71: Thanks for your comment.09:51
pave1patersonc: Would it be easy for you to say download big tar archive then unpack it?09:51
punitMaybe a bit off-topic, but why are CIP core images needed for CIP kernel testing?09:52
patersoncpave1: Internet connection to the boards under test isn't guarantied, hence the need to pre-install dependencies in the rfs. It is possible to do something like that though, if the tar archive is present in the repository that the test case is stored in.09:52
patersoncpunit: Technically they aren't. It just makes sense to me to use our own filesystem for testing, as it'll also help test cip-core.09:53
patersoncMaybe it's less essential than I think though...?09:53
pave1patersonc: That might be a good start. Actually even unpacking single tarball multiple times (so that total ammount of data is comparable to ammount of RAM) would be interesting.09:54
pave1patersonc: Is root filesystem for a test local or on NFS?09:54
patersoncNFS09:54
punitIt does, but then we run into two issues - 1. not having a wide test coverage as cip-core is expected to be minimal. 2. expanding cip-core to be a lot more than it is today to cover all the requirements of the various test suites09:54
kazu71punit Hello! My understanding is that it's better to use the same user land as CIP Core to test kernel because CIP kernel & core assume to be used together.09:55
kazu71by users of CIP09:55
patersoncI'm sorry everyone, I have to head to another meeting. I should be back in an hour or so for any follow ups if anyone is still online.09:56
pave1paterson: Ok..09:56
kazu71patersonc Thank you for the discussion!09:56
masashi910patersonc: see you.09:56
punitkazu71: I understand about them being tested together. But the kernel test surface is quite large. And if we rely on external test suites then the dependencies grow quite large09:56
pave1paterson: Thank you!09:56
kazu71punit One important thing is that providing packages for testing doesn't mean that CIP Core will be responsible to maintain them09:57
punitExactly, that's the point I wanted to get across09:58
fujita3Technically we can make another list for testing kernel, but I don't think it is necessary.09:59
punitSo then we are talking about assembling the necessary packages (including maintained) in a suitable image to enable testing09:59
kazu71punit Core should provide two package list: 1) the important packages that will be maintained, 2) Available package list (for testing, debugging, performance evaluation, etc.)09:59
kazu71fujita3 I have the same opinion10:00
punitkazu71: Makes sense. For 2) I wonder if we can defer to external distro such as debian...10:00
punitAlso the related question about should this also be available for tiny profile..10:01
*** kazu has joined #cip10:03
kazusorry, disconnected suddenly...10:04
kazupunit I have similar concerns about tiny because we need to implement recipes in meta-debian if the packages are not available yet.10:05
punit:)10:05
kazuActually, in generic profile, I guess we can install all Debian packages (in main component) by adding one or two definitions in kas file10:05
*** kazu71 has quit IRC10:07
punitYes, so we need to better understand how to test the filesystem themselves10:08
*** kazu has quit IRC10:13
*** pave1 has quit IRC10:19
* patersonc returns10:25
*** masashi910 has quit IRC10:27
*** fujita3 has quit IRC11:02
*** rajm has quit IRC11:21
*** rajm has joined #cip11:27
*** rajm has quit IRC11:33
*** rajm has joined #cip11:43
*** masashi910 has joined #cip12:50
*** masashi910 has quit IRC13:18
*** therisen has quit IRC16:18
*** cttpollard has quit IRC16:58
*** rajm has quit IRC22:52

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!