IRC logs for #cip for Thursday, 2019-01-10

*** tpollard has quit IRC07:18
*** rajm has joined #cip07:52
*** mungaip has joined #cip07:56
*** sangorrin has joined #cip08:12
*** tpollard has joined #cip08:19
*** vidda has joined #cip08:48
*** toscalix has joined #cip08:54
szlin#roll call09:00
szlinplease say hi if you're here09:00
szlinhi ._./~\09:00
szlin#topic kernel maintenance updates09:01
szlinBen imported more data in cip-kernel-sec09:01
szlin#action Daniel will start to work on v4.4.166-cip29-rt21 this friday09:01
szliniwamatsu: do you have any updates?09:01
bwhHi, sorry I'm late09:02
szlinbwh: (wave09:02
iwamatsuszlin: No, I am testing 4.19.y now.09:02
szliniwamatsu: thanks09:03
szlinbwh: do you have any updates? I've mentioned the cip-kernel-sec update09:03
bwhNo, I haven't done much else as I was mostly on holiday09:04
szlinFeel free to inform us if you need any collaboration work with cip-kernel-sec09:04
szlinbwh: thanks.09:04
szlinwagi: are you around?09:05
szlinany other points?09:05
szlin#topic kernel testing09:06
patersoncWe've been having some issues with the LAVA master. Restoring backups keeps failing. LF is currently investigating.09:06
patersoncNot a lot else to report due to the holidays09:07
szlinpatersonc: thanks.09:08
szliniwamatsu: could you elaborate 4.19.y testing?09:08
iwamatsuszlin: I am build, boot and LTP test on RZ board and QEMU.09:09
szlinany issues founded or new test cases added with new feature in 4.19.y?09:09
iwamatsuAnd I am updating cip-kernel-config to 4.19.09:10
szliniwamatsu: thanks.09:11
iwamatsuThere is no problem for now.09:11
iwamatsuPlease let me know if there is any valid document about test cases.09:11
patersonciwamatsu: I'm working to publish some test cases for the RZ board(s).09:12
patersoncMainly these will be designed to run in the LAVA env, but there is no reason not to run them directly on the board.09:12
sangorrinpatersonc: are you booting from NFS, NBDROOT or just from a ramdisk?09:13
sangorrinoh probably its written in the guide i have to read ^^09:13
patersoncAt the moment either NFS or ramdisk09:13
sangorrini had problems running ltp on ramdisk09:14
sangorrini think NFS has problems too09:14
iwamatsupatersonc: Do you have a document that I see about it?09:14
sangorrinin AGL they use nbdroot, or sth like that09:14
patersoncI haven't tried running LTS yet.09:14
patersonciwamatsu: Not yet. Will publish something when the scripts are ready.09:14
iwamatsupatersonc: I see.09:15
szlinpatersonc: the docs and scripts will be uploaded to cip-project right?09:16
szlin <-09:17
szlinpatersonc: great, we can start to share the test cases09:17
szlinany other points?09:17
patersoncNot from me09:18
szlin#topic CIP Core09:18
sangorrinJan prepared an ISAR version of cip core, its on the playground09:18
sangorrinwe have to integrate it on gitlab09:18
sangorrinwe decided to use gitlab groups and use separate repositories for each implementation09:19
sangorrinsome instructions and manuals will need to be updated09:19
sangorrinor just announce it on the list09:19
sangorrinpatersonc: are renesas manuals depending on the cip-core links?09:20
patersoncsangorrin: I don't think so, no09:20
sangorrinother than that, I am waiting for Siemens to fill my survesy ;_;09:20
sangorrinok, no more points from me09:21
szlinsangorrin: IIRC, Wolfgang mentioned they will fill out the survey recently on Monday09:22
szlinsangorrin:  BTW, helmut is looking for Kazuhiro. I think he wants to ask debian-cross stuff09:23
sangorrini will tell him, he is on my left09:23
szlinhelmut: are you around?09:23
szlinany other points?09:23
szlin#topic Software update09:24
sangorrini sent the survey09:24
sangorrini am thinking about using binary diffs and a A-B approach09:24
sangorrinbinary diffs: because iot devices usually run on sim cards that have a limit per month09:25
sangorrinor similar restrictions09:25
sangorrinA-B approach: because it causes lower downtime09:25
sangorrinwe can install on partition b in the background and then just reboot09:25
sangorrinill finish a draft of the architecture soon and sent it to the dev list for feedback09:26
sangorrinif you have any input please let me know anytime09:26
szlinsangorrin: Do you think it necessary to use live patch?09:27
sangorrinbtw: there are two types of binary diffs: sector based and file-based. I think I want to use file-based but not sure the implications09:27
sangorrinszlin: do you think that's stable?09:28
sangorrinI prefer a more conservative approach09:28
szlinsangorrin: of course not.09:28
sangorrinoh and about the bootloader.. i think updating the bootloader can be complicated sometimes09:28
sangorrinwe could brick our devices ;_;09:29
szlinperhaps A/B bootloader09:29
bwhszlin: Live patching is stable maintenance on the hardest difficulty level :-)09:29
szlinbwh: totally agree09:30
sangorrinhaha then I will avoid it :D09:30
patersoncA/B bootloader is really dependant on the device09:30
patersoncAnd where the flash can be accessed from.09:31
patersoncSome newer devices don't allow access to the secure boot medium from the unsecure/Linux world09:31
sangorrinpatersonc: good point, we also need to think about such devices with trustzone09:31
szlinand there're some security issues should be handled if we want to upload bootloder in OS.09:32
sangorrinmmm lets start with something simple09:32
fujita[m]sangorin: does A-B approach need double storage size?09:33
fujita[m]embedded manufacturers do not want to add extra area09:33
szlinfujita[m]: that's kind of trade-off09:34
sangorrinfujita: yes, that's why i did the survey09:34
sangorrinit seems that most members have enough space09:35
fujita[m]I'll check again the survey seet09:35
fujita[m]our customers always sais they want to reduce their storage09:36
sangorrinbut i think many devices cant afford it09:36
sangorrinfor example, routers or something with a small flash memory09:36
szlinfujita[m]: I can understand some low-level devices equipped very little size of flash09:36
sangorrinfujita: thanks for the feedback, another approach is use meta-updater which does not require a B partitions09:37
bwhI think this is going to a little too much detail for this meeting...09:37
fujita[m]thank you!09:37
sangorrinyeah, lets move the discussion to the list09:37
szlinany other points?09:38
sangorrinnothing else from my side szlin09:38
szlin#topic AOB09:38
szlinany other business?09:39
szlinthanks all09:39
fujita[m]iwamatsu: We're looking forward to the release tomorrow09:40
sangorrinthanks szlin!09:40
sangorrinsorry if my part was too long09:40
iwamatsufujita: Thanks!09:40
patersoncNo worries. It's good to have these discussions :)09:41
iwamatsuszlin: thanks you.09:41
*** sangorrin has quit IRC09:42
*** vidda has quit IRC09:48
*** rajm has quit IRC10:20
*** rajm has joined #cip10:43
*** rajm has quit IRC13:34
*** rajm has joined #cip15:23
*** iwamatsu_ has quit IRC16:10
helmutszlin: I'm sometimes around and read backlog when highlighted. I usually follow up within 24h.16:27
*** mungaip has quit IRC17:53
*** tpollard has quit IRC18:10
*** toscalix has quit IRC20:22
*** rajm has quit IRC22:52

Generated by 2.15.3 by Marius Gedminas - find it at!